+ All Categories
Home > Documents > Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5...

Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5...

Date post: 24-Aug-2020
Category:
Upload: others
View: 0 times
Download: 0 times
Share this document with a friend
51
Alex Yakyma PACIFIC EXPRESS
Transcript
Page 1: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

Alex  Yakyma                        

PACIFIC  EXPRESS        

           

Page 2: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   2  

ABOUT  THE  AUTHOR    Alex   Yakyma   is   a   methodologist,   trainer,   and   consultant  who   has   been   applying   Lean   and   Agile   practices   in   the  software  industry  for  over  a  decade.    Alex   joined   Dean   Leffingwell   in   2006   and   since   then   has  been   actively  working   on   Scaled   Agile   Framework   and   its  numerous   implementations   in   the   field.  Alex’s  broad  prior  experience   as   a   program   manager   in   highly   distributed,  multi-­‐cultural   environments   allows   him   to   perfect   his  

clients’   operational   capabilities   at   the   program   level,   cross   program   coordination,  and  portfolio  strategy.    Contact  Alex  at  [email protected]    Follow  on  Twitter:  http://twitter.com/AlexYakyma                                                        ISBN  978-­‐1-­‐4951-­‐2790-­‐8    

Page 3: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   3  

FOREWORD  BY  DEAN  LEFFINGWELL    Building   really   big   software   systems   is   one   of   the   most  challenging  intellectual  endeavors  facing  today’s  businesses.  To   help   address   this   challenge,   we’ve   created   the   Scaled  Agile  Framework,   (SAFe),   a  knowledge  base  of  proven  best  practices  that  you  can  use  to  help  bring  the  benefits  of  Lean  and  Agile  development  to  your  business,  and  also  better  the  lives—not   just   of   the   users   for   whom   we   build   these  systems—but   also   of   the  millions   of   software   practitioners  

who  create  them.  We  do  this,  most  simply,  because  better  software  makes  the  world  a  better  place.      That  scales.    To  accomplish  this,  SAFe  is  based  on  three  primary  knowledge  pools,  Lean,  Systems  Thinking   and   Agile   development.   These   values   and   principles   power   SAFe   and  provide   the  bedrock   that  make   it   “safe”  and  effective.    SAFe   is   freely-­‐revealed  and  publicly-­‐facing,   so   that   you   can   begin   to   apply   it   immediately   in   your   business  context.      Freely   revealed-­‐yes,  magic-­‐no.   Implementing  SAFe   is  hard  work.  New  values,   new  ways  of  working,  and  a  willingness  to  engage  and  fully  empower  the  people  who  do  this  important  work  are  all  required.  And  there  is  no  hidden,  or  proprietary,  secret  sauce   that  unlocks   the  magic.   If   there   is  any  magic   to  be   found,   it   comes   from  the  hearts  and  minds  of  those  dedicated  people  for  whom  the  status  quo  is  not  ok;  those  who   face   the   burden   of   change,   some   positively,   some   negatively.   SAFe   is   a  documented  system,  but  it  is  people  who  do  the  work.    To   that   end,   Alex,   has   put   together   this  work   of   semi-­‐fiction,   focused   on   one   key  element   of   SAFe,   the   Agile   Release   Train.   Semi-­‐fiction   because   the   business   and  characters   are   fictional,   but   the   challenges,   vignettes,   opinions   and   behaviors  expressed  were  real.  His  goal   is   to  help  you  understand  SAFe,   inside  out,   from  the  standpoint   of   the   people   doing   the   work.   He   has   chosen   this   novella   as   an  experiment  to  see  if  we  can  better  communicate  why  and  how  it  works.    So   let’s   leave   the   SAFe  website,  with   its   practices,   figures,   artifacts,   activities   and  references,  and  enter  this  fictional  world,  even  if  just  for  an  hour  or  two.  Let’s  see  if  we  can  find  an  even  better  understanding  of  why  SAFe  works  the  way  it  does,  and  perhaps   find  a   little   fictional  enjoyment   in   the  process.     I  know  I  did,  and   I  expect  you  will  too.      

Page 4: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   4  

I.  THE  BLACK  HOLE    “…A   hundred   and   ten…   approximately.   All   full   time.   This   program   grew   rapidly  during  the  last  year  –  from  some  forty  people  to  almost  three  times  that  size.  I  wish  our  processes  were  evolving  along  with  our  growth,  but  sometimes  it  seems  to  me  that  we   are   only   getting  worse   and  worse.”   The   lady   in   the   gray   elegant   suit   put  some  numbers  on  the  whiteboard  and  paused  for  a  few  moments.  “A  lot  depends  on  these  people,”  she  gestured  to  the  board,  “and  yet  this  program  is  like  a  black  hole  –  requirements   are   fed   in   but   nothing   gets   out.  Our   key   stakeholders   are   becoming  impatient   especially   because…”   she   got   stuck   for   a   few   more   moments,   perhaps  struggling   to   frame   her   statement,   “…because   we   are   Agile…   at   least   so   we  advertise.”    Some  people   in   the  room  grimaced  at   these  words,  others  smiled  and  some  chose  not  to  burden  themselves  with  any  reaction  just  yet,  probably  anticipating  more  fun  to  come.  Even  though  I  don’t  particularly  enjoy  other  people’s  awkward  moments,  these   can   actually   be   quite   useful   –   they   tell   you  what   your   customer   really   feels  more  accurately  than  a  thousands  words.  For  a  consultant  paying  attention  to  such  things   is   absolutely   vital.  We   just   got   started   and   things   are   already   popping   up.  Stephanie  seems  to  be  straightforward  enough  –  that’s  going  to  make  this  easier  for  all  of  us.  Whether   it’s   increasing  executive  pressure  she’s  under  (it  was  noticeable  even  during  our  first  phone  call)  or  just  her  personality,  I  think  I   like  the  direction  this  is  heading.  She  doesn’t  seem  to  be  overly  invested  in  their  current  methods  and  tools   either.   Maybe   she’s   just   one   of   those   heads   of   PMO   that   are   deeply   critical  about  the  results,  despite  all  the  politics  and  organizational  inertia.  So  far,  so  good,  but  I  need  more  detail:    “And   could   you  please   describe   in  more   detail  what   exactly   ‘Agile’  means   to   your  teams?  –”    “We  work  in  Sprints,  we  use  story  points,  we  do  all  that  Sprint  planning  stuff…  We  just  don’t  get  anything  out  of  it  in  the  end,”  a  gentleman  that  was  sitting  the  furthest  from  the  whiteboard  interjected.      “Ryan,   this   is   Josh,  one  of   the  product  managers  working  with   this  program,”   said  Stephanie   apologetically.   “We  used   to   call   this   role   ‘business   analyst’   but  with   the  adoption   of   Agile,   we   decided   to   make   a   switch,   because…”   she   rolled   her   eyes  desperately  trying  to  remember,  “Well,  I  guess  nobody  remembers  exactly  why.”        Meanwhile,  Josh  went  on…      “While  I  told  you  what  is  happening  process-­‐wise,  I  didn’t  say  that  I  see  any  value  in  that  process,”  he  clicked  his  pen  a  couple  of  times  and  added:  “as  she  said  –  stuff  gets  in  but  never  gets  out.”    

Page 5: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   5  

This  is  getting  more  and  more  interesting.  “Josh,  if  they  are  Sprinting,  don’t  they  also  do  Sprint  demo  in  the  end?”  I  asked.      Josh  only  chuckled  in  response,  grabbed  his  pen  again  and  got  back  to  clicking  it.      “I’m  Saheli,  I’m  a  product  manager  within  this  group,  like  Josh,”  a  lady  next  to  Josh  introduced  herself  with  a  light  charming  accent.  “Yes,  teams  do  demos  and  Josh  and  I   were   attending   those   for   quite   a   while,   but   all   they   show   is   mostly   unfinished  stuff…   Just   some   functionality   that   makes   little   sense   without   the   other  components.”    “Eventually   I   said   I’m   not   going   to  waste   any  more   time   on   those   demos   and  we  stopped  attending,”  Josh  jumped  in  again.  “If  things  keep  going  like  this,  I  think  we  will  all  have  to  start  looking  at  updating  our  résumés.  Personally  I  will  include  a  year  of  hyper-­‐Agile   experience  based  on   this,”  he   said   and   laughed.  Nobody  else   in   the  room,  however,  seemed  to  find  it  as  funny.        “That’s  why  we  contacted  you,  Ryan.  We  read  about  this  concept  of  the  Agile  Release  Train   and   we   thought   we   could   apply   it   to   this   program.   I   talked   to   my   former  colleagues  who   are  now   running   local  Agile   community   and  one  of   them   strongly  recommended   you.   She   actually   said   that   you’ve   been   through   some   really   tough  cases  in  your  consulting  career,  which  makes  me  believe  that  you  could  help  us  too.  She  mentioned  that  she  met  you  when  you  both  attended  a  conference   in  Orlando  last  year.”    I   can’t   figure   out   to   whom   she   could   possibly   be   referring,   but   that’s   not   what’s  important  right  at  this  moment.      “I   bet   those   other   cases   weren’t   as   tough   as   our   own   private   hell,”   Josh   noted  sarcastically,  now  busily  fidgeting  with  his  pen.      “My  understanding  is  that  the  whole  notion  of  the  Train  revolves  around  delivering  visible,   tangible   value;   and   I   think   that’s   exactly   what   we   need,”   said   Stephanie  getting  us  back  on  track.  “Tell  us  what  we  need  to  do.  What  does  the  transformation  process  look  like?”    “We  call  it  a  Quickstart,”  I  said.  “The  primary  action  happens  over  the  course  of  one  week   and   is   normally   preceded   by   careful   preparation   and   then   some   follow   up.  During   that   week,   I   will   work  with   teams   to   re-­‐align   them   to   a   common   process  model,   that’s   the   first   two  days.  Then,   for  the  next  two  days,  we  will  plan  our   first  Program  Increment:  PI   for  short.  Roughly  speaking,  you  can  think  of   it  as  a  release  planning   session.   And   finally,   the   last   day   can   be   used   for   Product   Owners   and  Scrum  Masters   to   dig   deeper   into   the   advanced   topics   of   how   to   operate   at   scale.  The  Quickstart  is  just  a  kick-­‐off  for  you  guys,  but  should  provide  enough  momentum  to  drive  the  transformation.”      

Page 6: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   6  

“How  long  does  the  preparation  stage  take?”  wondered  Stephanie.      “Typically  a  few  weeks.  We  need  to  ensure  that  we  are  all  aligned  on  key  priorities  and  that  we  can  actually  implement  them.  And  let’s  not  forget  all  the  logistics  for  the  Program  Increment  planning.”    “Logistics?  Like  what?”  she  asked.      “Well,   it’s   a   two-­‐day   planning   session.   Fully   collocated,   highly   collaborative   and  exclusively  focused  on  exploring  the  next  PI  scope  –”    “We   always   plan   releases   collocated.   We   are   kind   of   lucky   that   way.   The   whole  program  is  located  in  the  same  building…”    “That’s  great.  But  who  participates  in  it?”  I  asked.    “Well,  our  product  management,  who  you  already  know.  Engineering  management  for   the  program,”  she  pointed   towards   four  guys  sitting  right  next   to  me,   “and   the  Product  Owners,  who  aren’t  with  us  today.  That  makes  nine  more  people.”    “How  about  developers,  testers?...”    “What  do  you  mean?”    “Do  you  include  them  too?”    Stephanie   looked   surprised   and  puzzled.   “No…  What   do   you  mean,   include   them?  We  can’t  include  the  entire  program  into  the  planning…”    “Of  course  you  can.  Not  only  can  you,  but  you  should.  How  else  do  you  think  they  are  going  to  find  out  what  they  are  supposed  to  build  within  the  next  observable  time  frame?”    “The  Product  Owners  will  tell  them.”    Josh  burst  out  in  laughter.  “That’s  the  ‘telephone  game’  for  sure…”      “Are  you  serious?”  Stephanie  stared  at  me,  still  puzzled  and  clearly  ignoring  further  comments   from   Josh.   “Are   you   seriously   suggesting   that   we   put   everyone   in   the  same  room  and  have  them  plan  together?  All  hundred  and  ten  people?”    “May   I   ask   you,   Stephanie,   how   do   those   110   people   understand   and   manage  dependencies  today?”    “Well…”    

Page 7: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   7  

“They   don’t,”   interrupted   Josh.   “They   pretend   that   dependencies   don’t   exist.   They  each  iterate  in  a  complete  vacuum,  every  single  team.  And  that’s  why  when  the  time  comes   to   show   Saheli   and  me   their   current   progress,   all   they   have   is   a   bunch   of  disjointed,  esoteric  stuff.”    That  was   rough   but   helpful.   Actually,   despite   his   tremendously   sarcastic   attitude,  Josh   seems   to   understand   the   real   problem   here.   Stephanie   begins   looking  increasingly  puzzled.    I   outlined   the   plan:   “We   will   invite   all   the   teams   and   have   Product   Management  speak   directly   to   them,   presenting   the   vision   and   key   program  priorities.  We  will  then  let  the  teams  figure  out  how  much  of  that  they  can  realistically  deliver  within  the  Program  Increment.  They  will   sort  out   the  dependencies  and  ensure   that   they  can   provide   what   they   need   for   each   other.   They   will   address   significant   risks  together   to  clear   the  way   for  realistic  release  commitment.  Then  they  will  present  their  plans  back   to   the  stakeholders,   to  make  sure   that   the   intent   is   right,  and   the  program  is  ready  to  go  full  speed  in  the  right  direction  for  the  next  PI.”    Josh  stopped  his  clicking.  “Stephanie,  be  open-­‐minded…  It’ll  work  like  a  charm.”    It  was  really  hard  to  say  whether  he  was  teasing  or  being  completely  serious,  but  it  did  not  matter   to   Stephanie.   She  was   silently  processing  all   she  had  heard.   It  was  obvious  that  she  had  begun  connecting  the  dots.      “So,   that’s  why   the   article   I   read   about   the   Agile   Release   Train  was   talking   about  alignment  in  almost  every  other  sentence…”  she  said.    “Of  course.  Think  about   it.   If  a  group  of  110  people  are  busily  building  something,  but  never  able   to  produce  a  really  meaningful   increment  of  value  end-­‐to-­‐end,  how  much  of  their  work  is  a  waste?”    Stephanie  silently  nodded.      I  continued:  “Think  about  developers  and  testers  in  particular  –  people  that  directly  create  value   in   this  program.  They  are  supposed   to  make  an  enormous  number  of  decisions  at  the  micro  level  every  day.  For  example,  how  to  write  this  ‘IF  statement’  or  what  parameters  to  use  in  a  test  scenario,  or  what  interface  to  select  between  two  components   –   these   are   the   questions   for   which   they   badly   need   meaningful  context.  They  need  to  hear  from  the  business  owners,  the  product  management,  the  architects,  and  the  infrastructure  people  which  direction  the  Train  is  moving.  They  need   to   figure  out  between   themselves  –   the   teams  –  what  and  how   to  build.  Yes,  you  are  right,  alignment  is  a  key  driver  behind  the  notion  of  the  Agile  Release  Train.  Global   alignment   is   favored   over   local   team   ‘excellence’,   and   the   release   planning  session  is  a  key  first  step  in  that  direction.”    

Page 8: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   8  

“You  know,  Ryan,  even  though  I  agree  with  all  of  this,”  she  sighed,  “I  can’t   imagine  how  we  sell  it  to  the  business  stakeholders  or  even  to  the  teams.  There  will  be  a  lot  of  resistance.  We’ve  already  lost  the  trust  of  the  business  stakeholders  and  now  we  are  asking  for  two  more  days  of  their  time?  Not  to  mention  that  the  teams  would  be  distracted  for  four  days?  That’s  a  tall  order.”    “I   understand,   but   I’ll   tell   you,   Stephanie,   this   is   a   pretty   simple   cost-­‐benefit  conversation.”    Stephanie  quickly  sat  down  next  to  me,  ready  to  take  notes.      “Well,   first   and   foremost,   in   order   to   justify   the   investment   of   time,   effort   and  money,  we  need  to  understand  what  the  company  can  expect  to  get  out  of  it,”  I  said.      “At   least   some   end-­‐to-­‐end   functionality,”   interrupted   Josh.   “We   need   to   start  building  stuff.  If  we  can  deliver,  we  can  capitalize  on  that  and  the  sooner  we  do  it  the  better…”    “Fair   enough.   Now,   Agile   Release   Train   is   designed   to   enable   you   to   build  production-­‐ready  increments  of  value  at   least  every  ten  weeks,  as  well  as  produce  end-­‐to-­‐end   working   software   as   a   measure   of   progress   at   least   once   every   two  weeks.”    Stephanie  was   taking   notes   in   complete   silence.   I   let   her   finish   –   there’s  more   to  come.    “Over   time,   Trains   normally   manage   to   deliver   value   more   and   more   frequently,  therefore  detaching  development   cadence   from  delivery   schedule.   In  other  words,  you  deliver  when  the  business  needs  it  rather  than  only  at  the  Program  Increment  boundaries,   which   is   usually   eight   to   twelve   weeks.   I   picked   ten   as   a   typical  example.”    “That   sounds   really   good.”   Stephanie   finished   writing   and   turned   to   me   again   –  ready,  obviously,  for  more  ideas.      “Collocated   planning   every   ten   weeks   drives   alignment   of   the   entire   Train   to   a  common  vision.  Backed  by   frequent  end-­‐to-­‐end  system   integration  and   fortnightly  system  demos,   it  will   allow   this  group   to  build  more  value.   It’s   simple   logic:   there  will  be  much  less  rework  because  teams  never  get  out  of  synch.”    “Nice…  Give  me  one  more,”  said  Stephanie.      “Emphasis   on   code   quality   and   architectural   concerns   will   help   us   sustain   high  program   velocity   in   the   long   term,   and  will   prevent   technical,   design   and   quality  debt  from  accumulating.  This  means  that  our  ability  to  add  value  will  be  sustained  over  time,  which  is  not  a  typical  picture  across  the  industry…  unfortunately.”  

Page 9: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   9  

 “It   will   be   fortunate   for   us,”   said   Saheli.   “If   we  manage   to   implement   all   you   are  saying.”    Josh  made   a   funny   face,   clearly   suggesting   that   it   would   not   be   an   easy   journey.  “Those   of   us,”   he   quickly   added   to   the   gesture,   “Who   survive   this   transformation,  will   never   be   the   same…”   While   the   development   managers   and   Saheli   were  cracking  up  at  Josh’s  joke,  Stephanie  stayed  right  on  topic.    “On   the   other   hand,”   she   acknowledged   sadly,   “Most   of   our   teams   are   quite  demotivated  by  the  way  things  are  now.  I  doubt  that  they  will  be  very  supportive  of  these  changes…  or  of  any  changes  at  all.  Forcing  them  is  something  I  would  like  us  to  avoid   at   all   cost.   Last   month   we   lost   two   of   our   Scrum   Masters   and   a   bunch   of  developers  –  people  don’t  feel  happy  about  their  work;  group  morale  is  very  low…”    “I  understand.  We  will  address  that  too.  This  model  creates  benefit  not  only  for  the  business,  but  is  built  on  a  strong  foundation  of  respect  for  the  people  who  actually  create  value.”    “Tell  me  more,”  said  Stephanie,  busily  writing  more  notes  on  a  new  sheet.      “Teams  will  no  longer  plan  blindly.  They  will  now  hear  program  priorities  directly  from   the  key  stakeholders.  And   they  won’t   just   ‘hear’,   it’s   a   two-­‐way  conversation  where  teams  can  and  should  ask  questions  directly  of  the  stakeholders  to  figure  out  what  and  how  they  are  building  in  this  PI.”    “Okay…”    “Next  –  their  focus  will  be  on  that  which  really  matters  –  end-­‐to-­‐end  value  delivery.  They  will  score  every  two  weeks.  Their  ultimate  goal  every  Sprint  will  be  to  deliver  a   slice   of   value   together:   something   they   can   proudly   demonstrate   to   the  stakeholders.  But  even  more  importantly,  they  will  now  know  themselves  that  they  are   making   real   progress.   That   will   bring   back   esprit   de   corps   and   make   the  workplace  an  enjoyable  place  to  be.”    “Beautiful…”  said  Stephanie,  expecting  at  least  one  more  item.      “We  will   foster  software  craftsmanship  and  make  sure  that  we,  as  a  program,  take  pride   in   delivering   quality   code.   No  more   just   ‘doing   it’.   Every   increment  we  will  build  quality   in.  We  will   therefore   understand  our   real   velocity   as   a   group   and  be  able  to  sustain  it  over  a  long  period  of  time.”    “I  like  it.  Whether  or  not  it  will  be  an  easy  sell  –  only  time  will  tell.  I  personally  think  it’ll   be   a   tough   sell.   Nor,   I’m   afraid,   will   the   implementation   be   easy   in   our  environment.   But   I   actually   have   authority   to   approve   the   budget   for   this  Quickstart.”  She  paused  a  moment.  “Last  question,  Ryan.  When  can  we  get  started?”  

Page 10: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   10  

II.  LAND  OF  THE  SLUMBERING  SUN    San  Diego  is  an  amazing  place,  especially  in  the  fall.  Nature  begins  to  cut  its  ties  with  summer  in  preparation  to  what  passes  for  winter  around  here.  The  temperature  is  perfect,  although  the  water  is  a  bit  colder  than  comfortable.  The  ocean  seems  a  little  grimmer,  but  I  suspect  I’m  the  only  one  who  cares  to  notice  that.  Even  having  visited  just  a  few  times,  I  feel  I  know  this  place  quite  well  and  really  enjoy  it  here.  It  would  be   nice   to   have  more   Quickstarts   on   the  West   Coast,   but   if   one   could   make   that  happen  every  time  –  wouldn’t  that  be  a  super  power?  Like  magic,  almost  on  par  with  making  software  development  really  efficient  at  large  scale.    Tomorrow  is  a  really  big  day  –  we  start  our   two-­‐day  release  planning  session  and  much  will  depend  on  how  that  goes.  The  preparation  took  two  weeks  –  a  very  busy  two  weeks  –  not  an  uncommon  thing  when  a  program  initiates  this  process  for  the  first   time.  And  not  everything  went  as  smoothly  as  expected…  as   if   it  ever  does  at  scale.      The   first   thing   we   did   was   to   agree   on   the   cadence.   From   the   start,   Stephanie  managed   to  quickly  gather  key  people  and   initiate  a  good  discussion.  They  agreed  on  a  ten-­‐week  duration  for  program  increment,  even  though  it  took  some  nerve  to  do  so.  Some  Scrum  Masters  were  overzealous  in  defending  their  existing  three-­‐week  Sprint  cadence,  which  in  no  way  adds  up  to  a  ten-­‐week  PI.  But  even  apart  from  that  arithmetical  glitch,  by  insisting  on  three  week  Sprints,  they  would  really  complicate  synchronization   across   the   entire  Train,   as   lots   of   teams  were   running   in   shorter,  two-­‐week   cycles.  Aligning   the   Sprint   cadence  makes   synchronization   substantially  simpler.   And   since   most   of   the   teams   were   used   to   operating   in   two-­‐week  timeboxes,  we  finally  managed  to  convince  the  rest  of  the  program  to  do  the  same.  In   cases   like   this   you   need   a   clear,   strong   argument   and   such   an   argument   was  presented  to  the  teams.  When  they  fail  to  deliver  end-­‐to-­‐end  value,  sometimes  one  wonders,  what   can   be  more   important   than   synchronization   across   teams?  When  key  stakeholders  claim  that  they  see  no  output  from  the  program,  it  is  hard  to  argue  and  defend  a   local   team  caprice.   Stephanie  did   a   lot   of   selling  during   this   session,  strengthening  my  faith  that  there  is  a  future  for  ‘Pacific  Express’  (she  came  up  with  this  name  for  the  Train).  With  this  basic  input  we  were  cleared  to  move  on  to  other  things.        The   next   big   problem  was   frequent   integration.  We   had   a   very   painful   discussion  involving  technical  experts  from  different  teams,  as  well  as  architects  and  some  test  engineers.  The  way  that  process  had  evolved  was  less  than  ideal  –  teams  were  each  working   in   their   own,   isolated   branch   of   code.  Did   they   have   a   common  program  branch?  Yes,  and   the   teams  actually  worked  under   the  assumption   that   they  were  regularly  synchronizing  with   the  program  branch.  The  resulting  problem  was   that  they  would   check   out   code   from   that   common   branch   into   their   own,   but   almost  never   check   anything   back   in.   So,   when   nobody   pushes   anything   into   a   shared  branch  of  code,  guess  what’s  in  it?  Sprint  after  Sprint,  almost  no  changes  occurred  in  

Page 11: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   11  

the   shared   codebase.   No   wonder   ‘synchronizing’   this   way   was   shockingly   easy:  nothing  was  happening!  As  soon  as  we   instituted  a  different  experiment  –  agreed-­‐upon  mandatory  daily  check-­‐in  of  code  –  the  flawless  Happy  Land  turned  suddenly  into  a  world  of  pain  and  suffering.      Luckily,   though,   Saheli   provided   tremendous   support.   She   emphasized   the  importance  of  the  end-­‐to-­‐end  increments  from  the  teams  every  two  weeks.  And  how  could  they  produce  those  without   frequently   integrating?  “I  couldn’t  even  begin  to  imagine”,  she  said.  Eventually  the  thread  was  resolved.   It  was  decided  to   integrate  across   the   entire   Train   at   least   three   times   per   Sprint.   These   integration   points  would   be   coordinated   carefully   on   a   Sprint-­‐by-­‐Sprint   basis,   assisted   by   a   newly  created   System   Team.   Their   primary   responsibilities   were   to   maintain   the  integration   and   testing   environments   and   assist   the   rest   of   the   program  with   the  integration  process.  Despite  the  fact  that  such  a  team  was  created,  Stephanie  made  it  clear   to   everyone   that   the   primary   responsibility   for   end-­‐to-­‐end   integration   lies  with   Agile   teams.   After   a   few   days   of   experimenting,   the   teams   also   agreed   on   a  critical  rule  –  the  Green  Build  Policy.  Stated  briefly:  if  integration  isn’t  successful  (i.e.  not   ‘green’),   then   everybody   on   the   Train   stops   and   swarms   around   the   problem  until  it  is  resolved.  The  implication  is  that  no  new  functionality  is  being  built  on  top  of  an  incorrect  one,  which  would  cause  a  much  bigger  problem  later.  Instead,  every  new  step  is  laid  on  top  of  fully  integrated  revision  of  code.      The   organizational   structure   did   change   a   little.   It   was   decided   to   reorganize   the  maintenance  work  such  that  the  maintenance  team  would  dissolve  and  in  its  place,  maintenance   engineers   would   join   other   Agile   teams.   This   change,   proposed   by  Stephanie,  was   a   smart  move,   in  my  opinion.   There  had   always  been   a   big   gap   in  knowledge  between  the  maintenance  team  and  the  rest  of  the  program,  because  the  other  teams  never  had  an  opportunity  to  do  an  efficient  knowledge  transfer.  On  the  other  hand,  maintenance  guys  were   complaining  all   the   time  about   system  design  that   did   not   foster   maintainability   of   code.   This   was   due   to   poor   handling   of  dependencies  between  classes,  bloated  methods  and  functions  almost  impossible  to  comprehend,   and   very   tight   coupling   throughout   the   codebase.   Meanwhile,   the  eleven-­‐member  Argonauts  team  was  split   in  two.  These  two  changes  left  the  Train  with  the  same  number  of  teams  –  fourteen.      The   next   big   undertaking   during   the   preparation   was   agreeing   on   program  priorities.   This   process  was   particularly   ugly   in   the   very   beginning.   It   turned   out  that   Josh  and  Saheli  didn’t  actually  maintain  a  single  program  backlog,  but   instead  worked  with  teams  directly,  providing  some  scope  of  work  to  the  product  owners.  The  first  time  we  brought  things  together  in  the  same  backlog,  it  spawned  a  flurry  of  emotional   arguments   between   the   two   about   whose   stuff   was   more   valuable.   At  some  point,  I  realized  that  they  were  simply  shouting  at  one  another.  Stephanie  had  to   calm   them  down,  which   she   turns  out   to  be  quite  good  at.  Eventually,   after   the  introduction  of  economic  prioritization  with  Weighted  Shortest  Job  First   (WSJF),   the  conversation   shifted   into   a  more   productive   phase.   The   core   idea   here   is   simple:  every  feature  in  the  program  backlog  has  a  certain  Cost  of  Delay,  or  CoD  for  short.  

Page 12: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   12  

CoD  stands   for   the  amount  of  unfulfilled  benefit  when  delivery   is  not  made   to   the  customer.   The   Cost   of   Delay   of   the   feature,   divided   by   the   duration   it   takes   to  implement,  provides  a  good  numerical  expression  of  economic  priority.  This  is  quite  intuitive,   because   the  higher   the   cost  of  delay,   the   sooner  you  want   to   get   started  with  that  feature.  And  conversely,  the  more  time  it  takes  to  implement  a  feature,  the  later  you  start  with  it.  Otherwise,  other  features,  even  smaller  ones,  will  have  to  wait  a  long  time  to  be  delivered.  We  used  a  very  specific  expression  for  Cost  of  Delay  that  reflects  different  aspects  of  a  feature  and  uses  feature  size  as  a  proxy  for  duration:    

WSJF  =  (Business  Value  +  Opportunity  Enablement  +  Time  Criticality)  /  Size    We   then   gave   Saheli   and   Josh   a   chance   to   rank   each  of   these   four   parameters   for  every   feature,   in   relative   numbers,   using   Fibonacci   sequence   (just   like   the   one  Scrum  teams  use  to  estimate  the  size  of  their  stories:  0,  1,  2,  3,  5,  8,  13,  …).  The  result  was  staggering:  they  managed  to  reach  reasonable  agreement  on  almost  all  features  except  one.  Enter  Stephanie’s  negotiation  skills  mixed  with  the  Lean  economics.  So,  we   ended   up  with   a   shared   understanding   and   agreement   on   key   priorities.   And  perhaps   most   importantly,   a   unified   program   backlog,   prioritized   by   economic  value.  With  that  accomplished,   the  Train  was  ready  to  pick  some  top  priorities   for  implementation.      The   preparation   for   the   remainder   of   the   areas,   including   the   logistics,   went  relatively   smoothly.  Nevertheless,   Stephanie   and   I  were   clearly   sensing  an  uneasy  atmosphere   across   the   board.   Key   program   stakeholders,  who   Josh   and   Saheli   de  facto  report  to,  agreed  to  participate,  but  did  so  very  reluctantly.  They  were  visibly  distant  in  all  conversations  regarding  the  transformation.  It  is  hard  to  establish  trust  when   there’s   nothing   to   deliver   –   I   can   understand   that.   One   of   the   stakeholders,  John,  even  said  that  he  would  let  us  play  with  the  planning  and  other  things  like  that,  but  that  he  had  no  faith  in  it.  That’s  embarrassing,  but  Stephanie  and  I  are  going  to  prove  that  it’s  the  real  deal,  or  fail  and  fail  big  tomorrow.      But   that’s   tomorrow.  Today   I   feel  a  need   to  simply  give  my  mind  a  break   from  all  these  thoughts  that  have  piled  up  over  the  last  few  days.  Right  after  finishing  the  last  session  with   Stephanie,   I   slipped   out   on  my   escape   route   –   La   Jolla   Village  Drive,  which   seems   to   take   me   to   another   world   in   a   matter   of   minutes.   I   hastened   to  disappear   in   the   evening   traffic…   for   I   know   the   reward.   Under   the   spell   of   the  comforting  whisper  of  the  ocean,  I  can  finally  stop  thinking  and  worrying  and  fully  fuse  with   the   tremendous  vastness  of   the  sea,  as   it  prepares   to  shelter   the  sun   for  the  night.  The  day   finally  surrenders   to   the  magnificent  closure  ceremony  and   the  waters   darken,   as   does   the   whisper   of   incomprehensibly  mysterious   tongues.   No  day  repeats   itself  –   tomorrow  will  be  another  stream  of  unique  moments,  most  of  which  the  world  will  carelessly  forget.            

Page 13: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   13  

III.  GETTING  THEM  BACK    ‘Town  Hall’  is  filled  with  the  humming  sound  of  over  a  hundred  people  getting  their  morning  coffee,  pulling   laptops  out  of  backpacks,  and  checking  smart  phones  –  all  the  usual   a.m.   procedures  which   you   can  do  half-­‐asleep.  At   7:50   a.m.,  most   of   the  chairs  in  the  room  are  already  occupied.  Fourteen  tables,  one  per  team,  take  up  most  of   the   room,   leaving   some   space   around   the   walls.   A   few  more   tables   have   been  brought  in  to  handle  stakeholders  and  other  participants  that  aren’t  among  the  Agile  Teams.   Stephanie   is  busily   explaining   something   to   the   IT   folks   in   the  back  of   the  room,  pointing  at  the  projectors  and  large  speakers  in  the  front.      A   few  stragglers  appear   in   the  doorway  and  quickly   join   their   teams.  Stephanie   is  ready  to  kick  this  off.  She  takes  a  quick  sip  of  coffee  and  taps  the  microphone  with  her   fingernail   a   couple   times,  which  apart   from  a   sound  check   function,   signals   to  everyone  on  the  Train  that  we  are  about  to  get  started.    “Good   morning   guys,”   she   says   in   a   very   cheerful   voice,   causing   an   avalanche   of  ‘good   mornings’   in   response.   “Earlier   today,   when   putting   team   name   tents   –   all  those  Wolverines  and  Argonauts  and  Spartans  and  what  not  –  on  every  table,”  she  smiled  and  paused  for  a  second,  “I  realized  that  we  were  missing  the  most  important  team   name.   We   are   all   one   team.   We   are   all   here   together   today,   on   our   first  Program   Increment   planning,   because   as   we   learned,   we   are   not   just   a   bunch   of  isolated   small   teams.   The   only   way  we   can   deliver   value   is   by   working   together.  Therefore,   I   would   like   us   all   to   remember   that   we   are   first   and   foremost   the  members   of   a  much   bigger   team…”   Early   in   the  morning   Stephanie   had   carefully  done  her  homework.  She  turned  now  to  the  easel  and  tore  off  the  top  page  of  the  flip  chart,  revealing  the  name  for  the  whole  program  –  ‘Pacific  Express’.              Stephanie  clearly  had  the  audience’s  attention.  She  went  on:      “Today  we  will  do  our  first  big  thing  as  a   larger  team…  as  an  Agile  Release  Train.”  She  pointed  at  me:  “We  have  Ryan  with  us  for  the  two  days  of  planning.  Many  of  you  already   know   him   from   the   preparation   workshops   and   the   team   training   on  Monday   and   Tuesday.   Just   to   remind   you   all,   Ryan   is   a   consultant,   trainer   and  enterprise   coach   who   is   helping   us   implement   Scaled   Agile   Framework   in   our  organization.  This  program   is  our   initial   foray   into  Agile   at   scale   and   is   extremely  critical  to  the  success  of  our  business.  Our  program  will  soon  begin  operating  as  an  Agile   Release   Train   and   will   gradually   learn   how   to   deliver   meaningful   value  together,  step  by  step.  At  this  point  I  would  like  Ryan  to  take  the  stage…”    “Thank   you,   Stephanie.”   I   am   a   little   nervous.   No   matter   how   many   times   I   go  through  the  PI  planning,  it’s  always  a  big  deal.  “Yes,  Stephanie’s  right.  Each  of  you  is  a  member   of   two   teams   –   your  Agile   team   and   the  Train.   The   notion   of   the  Agile  Release  Train  is  one  of  the  cornerstone  concepts  of  the  Scaled  Agile  Framework  (or  SAFe  for  short).  Many  of  you  have  already  heard  about  SAFe,  but  let’s  make  sure  we  

Page 14: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   14  

all  clearly  understand  what   it   is  and  why  we  care.  SAFe  is  a  model  of  a  Lean-­‐Agile  enterprise.  Agility  had  proved  to  be  extremely  efficient  for  small  teams,  but  as  soon  as  large  businesses  began  implementing  the  method,  it  became  apparent  that  Agility  itself  was  not  inherently  suited  to  scale.  Out  of  this  disconnect  the  need  grew  for  a  thinking  tool   to  operate   in  a   large-­‐scale  environment.  And  such  a  tool  was  already  out   there   –   Lean   thinking   –   a   flow-­‐oriented   systems   thinking   approach   that  originated  from  Toyota  Production  System,  which  many  of  you  have  heard  of.  Lean  allows  us  to  abstract  for  a  moment  from  how  to  develop  software  in  an  iterative  and  incremental   manner,   and   instead   concentrate   solely   on   the   flow   of   value.   This  radical   emphasis   on   the   flow   allows   us   to   apply   this   thinking   paradigm   to   the  organization  as  a  whole.  It  serves  as  guidance  when  we  want  to  scale  agility.  It  also  governs  the  work  of  Agile  teams  so  that  they  can  understand  and  deliver  value  in  a  larger   context.   Lean   suggests   a   couple   of   simple   ideas,   that   nevertheless  substantially  facilitate  the  flow  of  value:  operating  at  low  levels  of  work-­‐in-­‐process,  minimizing  the  queues  in  the  flow,  and  constantly  optimizing  the  system  as  a  whole  –   all   to   achieve   the   shortest   sustainable   cycle   time   throughout   the   organization.    SAFe  utilizes  Lean  thinking  to  scale  Agile  best  practices.  And  the  first  such  level  we  scale  to   is  program.   Just  as  the  Agile  team  is   the  smallest  group  of  people  that  can  define-­‐build-­‐test  work,  there  is  another  level  –  Agile  Release  Train,  which  is  capable  of  delivering  solutions.  This  is  a  key  building  block  for  organizations  that  depend  on  software   development   for   their   success.   If   you  would   ask  me   how   to   define  Agile  Release   Train   in   a   single,   short   sentence,   I   would   suggest   that   it   is   simply   a  continuous  flow  program.  All  we  do  as  a  Train  is  to  accelerate  value  delivery  to  the  business.   Lean   thinking   applied   to   scaling   agility   at   this   level   leads   us   to   the  following  rules  for  the  Agile  Release  Train:    

• Train  is  a  self-­‐organized  team  of  Agile  teams  (50–125  individuals)  that  plans,  commits,  and  executes  together.  

• Program   Increment   is   a   fixed   timebox   –   a   major   planning   and   alignment  horizon  for  the  Train  –  typically  four  to  six  Sprints.    

• Teams  have  synchronized  Sprint  boundaries  within  the  PI.  • Everyone  on  the  Train  is  aligned  to  a  common  mission  via  a  single  program  

backlog,  consisting  of  features  sequenced  in  order  of  economic  priorities.  • The  Train  operates  under  architectural  and  UX  guidance.  • Every  two  weeks,  the  Train  produces  fully   integrated  valuable  increment  of  

the  solution.”    If  we   do   everything   right   during   this   planning   session,  we  will   indeed  witness   an  amazing  metamorphosis  in  this  room  –  a  bunch  of  disjointed  teams  will  turn  into  a  qualitatively  distinct  construct  –  an  Agile  Release  Train.  It  was  important  to  set  the  context   and   explain   to   them   what   Train   and   SAFe   are,   but   they   will   really  understand  it  only  by  experiencing  it.  The  planning  session  is  the  magic  wand…  It  is  time  to  introduce  the  agenda  for  it.      

Page 15: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   15  

“PI  planning  is  a  two-­‐day  process.  The  goal  is  simple  –  to  get  aligned  on  key  program  priorities   and   understand  what  we,   as   a   Train,   can   realistically   deliver  within   the  next   ten   weeks.   When   we   say   ‘realistically’,   we   mean   that   dependencies   are  identified  and  resolved,  risks  managed  and  objectives  for  each  team  agreed  upon  by  the  business  owners.  Here  is  how  we  are  going  to  do  this  –”    “Sorry,  I  have  a  question,”  a  lady  from  the  Argonauts  team  raised  her  hand.  “Is  this  process  similar  to  the  Sprint  planning?”      “Yes  and  no.  Yes,  in  the  sense  that  PI  planning  is  to  the  Train  what  Sprint  planning  is  to   the   Agile   team.   At   the   same,   time   it   has   many   additional   steps   and   process  requirements  that  differentiate  it  from  the  Sprint  planning.  The  exact  difference  we  will  be  able  to  experience  in  just  a  little  bit.    “The  first  half  of  the  day  today  will  consist  of  various  briefings.  Business  owners  will  present   the   overall   direction   for   the   business.   Product   Management   will   then  present   the   solution   vision   and   the   top   features   to   be  built   in   the  PI.   Some  of   the  Product   Owners   already   have   some   seeded   stories   –  we   did   initial   breakdown   of  features  into  stories  a  few  days  ago.  It  was  a  useful  exercise,  but  please  keep  in  mind  that  you  may  change  the  breakdown  as  you  see  fit  as  long  as  your  product  owner  is  onboard  with  it.  The  goal  is  to  drive  maximum  value  rather  than  stick  to  the  initial  breakdown  of  features.    “Next   we   will   have   the   overview   of   architecture   and   engineering   practices.  Development  infrastructure  will  also  be  discussed  as  part  of  this  briefing.  We  need  to  know  not  only  what  we  are  supposed  to  implement  as  a  Train,  but  also  how  we  are  going  to  do  that  implementation.  Again,  this  is  not  a  detailed  system  design;  it’s  a  high-­‐level   guidance.   This   concludes   the   set   of   briefings   that   provide   the   required  planning  context  for  us.      “Next  up  will  be  the  Team  Breakout  session.  This   is  where  the  teams  actually  plan  the  PI.  You  guys  will  have  to  break  down  features  into  stories  and  put  them  in  your  PI  plan.  While  we  will  discuss  in  detail  what  those  planning  artifacts  are  and  how  to  effectively  use  them,  let  me  make  it  clear  from  the  beginning  –  the  key  output  of  the  planning  for  each  team  will  be  a  set  of  Team  PI  Objectives  –  concise  statements  that  are  meaningful  and  valuable  to  the  business  owners.  Everything  else,  even  stories,  are   just   tools   to  arrive  at   a   succinct,   executable   set  of  objectives  –   let’s   remember  this  throughout  the  course  of  the  planning  –”    A  young  guy  from  the  ‘A-­‐Team’  raised  his  hand:  “How  are  objectives  different  from  features?  Or  they  are  the  same  thing?”    “Good   question.   In  many   cases,   objectives  will   actually   be   features.   But   that’s   not  always   the   case.   Think   of   some  milestones   like   supporting   user-­‐testing   event,   for  instance.  It’s  not  a  feature  per  se,  but  makes  for  a  meaningful  PI  objective  that  has  business   value.   Another   example   could   be   a   significant   research   effort.   The  

Page 16: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   16  

functionality   itself,   for  which   the   research   is   being   planned,  will   be   part   of   future  PIs;  or  it  may  not  be  committed  at  all  if  the  research  proves  to  be  negative.  But  the  research   itself   is   a   valuable   objective   within   the   PI.   Furthermore,   many   features  require  more  than  one  teams’  participation.  In  this  case,  each  team  will  have  some  objectives  that  may  contribute  to  the  feature  but  not  entirely  complete  it.  You  may  discover  more  examples   later   today   just  by  walking  around   the  room  and  reading  what  other  teams  have  on  their  objective  sheets.      “But  to  summarize,  I  can  say  that  features  are  the  initial  input  to  the  planning,  while  team  PI  objectives  represent  the  output.  The  purpose  of  the  PI  planning  session  is  to  conduct   some   sort   of   reality   check   and   to   understand   what   the   ideal   set   of   top  features  means   from   the   team   perspective.   This   is   possible   once   they   thoroughly  comprehend   the   implementation   and  possible   risks.   It   now  becomes  obvious  why  we   break   down   features   into   stories.   It   is   an   analysis-­‐synthesis   process.   First  (analysis)   teams   break   it   down   into   smaller,   actionable   and  more   understandable  items.  Then  they  try  to  understand  each  other’s  dependencies,  identify  impediments  and   risks,   re-­‐scope,   if   needed,   and   plan   for   supporting   activities   like   research,  maintenance,  etc.  Once  reality  has  set  in,  they  ‘integrate’  it  back  to  the  same  level  of  abstraction,  basically  by  summarizing  those  detailed  plans  via  the  team  PI  objectives  (synthesis).  That’s  how  we  should  view  this  process.      “I  therefore  ask  you  to  do  the  following:  every  team  must  at  all  cost  get  to  the  draft  PI   objectives   by   the   end   of   the   day   today,   otherwise   neither   the   teams   nor   the  business   owners   will   be   able   to   understand   where   we   are   with   respect   to   the  planning  process.  It   is  better  to  have  a  very  rough  plan  of  all   five  Sprints  in  the  PI,  and  thus  be  able  to  derive  the  PI  objectives,  than  to  have  only  two  or  three  Sprints  planned  very  accurately,  and  still  have  no  understanding  of  the  overall  PI  outcomes.  In   other   words,   don’t   get   stuck   on   too   much   detail   today.   I   especially   encourage  Scrum  Master  to  very  carefully  facilitate  the  process.  Identify  planning  impediments  as  early  as  possible  and  make  sure  the  teams  make  a  shallow  pass  over  the  entire  PI  scope,  rather  than  get  lost  in  the  detail  and  end  up  with  nothing  today.  Let’s  keep  in  mind,  that  we  will  have  the  entire  day  tomorrow  to  get  deeper  into  the  nuances  of  the   functionality   and   other   concerns.   To   help   teams   stay   on   track,   we   will   have  Scrum  Master  check-­‐in  every  45  minutes.  I  will  be  facilitating  those  meetings  while  the  rest  of  the  team  members  continue  working  on  their  plans.    “Up   next   will   be   the   Draft   Plan   Review,   where   every   team   will   present   their   PI  objectives,   as  well   as   rough  description  of   the   flow  of   the  Sprints:  what  gets  done  and  when.  This  is  a  short  and  focused  presentation,  typically  two-­‐to-­‐three  minutes,  immediately  followed  by  a  Q&A  from  the  business  owners  and  other  teams.  This  will  conclude   the   day   for  most   of   us,  while   program   stakeholders,   Scrum  Masters   and  Product  Owners  will  stay  for  the  Management  Review  and  Problem  Solving  meeting.  We   will   go   through   what   we   learned   from   the   team   plans   and   if   any   corrective  actions  are   required,   this   is   the   time   to  make  such  decisions.  Scope   trade-­‐offs  and  even  changes  to  the  team  composition  can  be  made  at  this  point.    

Page 17: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   17  

“We  will  begin  the  next  day  by  presenting  these  adjustments  to  the  rest  of  the  Train,  after  which   there  will   be   another   three   hours   for   team  breakout.   This   is   the   time  when  PI  plans  are  further  refined  and  business  owners  assign  business  value  to  the  objectives   for   each   team.   Once   teams   are   finished   planning   PI   scope,   they   will  present  final  plans  to  the  business  owners.  After  this  we  will  have  a  joint  session  for  risk  management   and,   finally,   teams  will   have   an   opportunity   to   do   a   confidence  vote.”    After  my  introduction  to  the  process,  Stephanie  invited  the  first  presenters:  John  –  Head   of   Retail   Automation,   and   Tanya   –   VP   Development.   The   impact   this  presentation  made   on   the   audience  was   tremendous.   In   fact,   it  was   the   first   time  most  of  the  team  members  had  ever  heard  from  someone  on  the  business  side  and  the  information  John  presented  provided  invaluable  context  for  the  whole  Train.      

   He   discussed   the   current   state   of   the   retail   business.   Then   he   talked   about   the  company’s   substantial   dilemma  of   brick-­‐and-­‐mortar   versus   ecommerce   as  well   as  business   automation   trade-­‐offs   associated  with  both  paths.   John  also   talked  about  the  key  strategic  theme  of  expanding  the  number  of  ways  our  customers  can  select  products,   buy   them,   and   have   them   delivered.   Customer   base   segmentation   and  better,  more   focused  outreach  to   the  segments  with  a  particular  value  proposition  was  another  big  theme.  Although  he  was  using  a  microphone,  the  audience  was  so  quiet  you  could  hear  a  pin  drop.  After  he  finished,  Stephanie  asked  if  there  were  any  questions   –   the   audience   remained   silent,   still   digesting   a   boatload   of   valuable  information.      You  may  wonder  how  it  is  that  the  teams,  working  in  the  same  organization  as  John,  were  unaware  of  the  most  basic  business  context  for  their  priorities?  It  should  come  

Page 18: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   18  

as   no   surprise,   given   that   the   methods   we   used   for   software   development   for  decades   were   built   for   anything   but   alignment.   Disjointed   activities   in   a   largely  silo’ed   organizational   structure   had   historically   fostered   sub-­‐optimization   of   the  functions,  but  not  the  flow  as  a  whole.  A  simple  thing  that  accomplishes  miracles  –  face-­‐to-­‐face  communication  –  was  unimaginably  far  off,  replaced  instead  by  tons  of  documentation  and  a  plethora  of  isolated  business  analysis  methods  and  tools.  With  the   advent   of   Agility,   the   industry   got   a   glimpse   of   hope   as   the  Manifesto   clearly  called   for   ‘face-­‐to-­‐face   communication’   and   ‘business   and   development   working  together’.   But   something   else   happened   in  most   large   organizations:   for   many   of  them,  even  with  the  adoption  of  the  new  roles  and  rules,  both  ‘Product  Owners’  and  their  teams  still  stood  far  off  from  any  real  business  context,  or  the  user.      We   as   an   industry   had   failed   again   because   of   our   natural   propensity   to   value  methods  and  practices  over  process  goals  and  systems  thinking.  Picking  a  developer  and   calling   him   a   ‘Product   Owner’   does   not   help   bridge   the   gap   between  development  and   the  business.  Having  daily  standups  does  not  create  any   face-­‐to-­‐face   communication   between   those   who   understand   what   needs   to   be   built,   and  those  who  are   supposed   to  build   it.   PI   planning   in   SAFe  has   a   secret  weapon,   the  likes  of  which  is  hard  to  find…  probably  because  it  took  a  genius  to  figure  out  that  simply  putting  people  together  in  the  same  big  room  and  having  them  talk  to  each  other  helps  them  stay  on  the  same  page.            Tanya’s  presentation  was  also  helpful.  She  provided  good  insight  to  the  technology  advancements.  She  introduced  an  important  theme:  shrinking  the  technology  stack  that  had  grown  uncontrollably  over  the  last  three  years,  ending  up  in  a  hairy  ball  of  different   tools,   libraries,   frameworks   and  platforms.  Many   of   these   duplicate   each  other   and   came   into   existence   as   a   matter   of   personal   preferences   of   different  teams,  as  opposed  to  any  rationale.  The  big  theme  for  this  PI,  as  she  noted,  would  be  to  move   all   existing   VB.NET  modules   to   C#,   as   they   are  much   easier   to  maintain.  Reducing  the  number  of  different  configuration  management  tools  used  for  different  aspects  of  data  management,  deployment  and  production  monitoring  was  another  vector  for  this  PI.      I  was   surprised   however,  when   after   finishing   their   presentations,   both   John   and  Tanya  grabbed  laptops  and  made  their  way  towards  the  door.  Puzzled,  I  glanced  at  Stephanie  but  she  didn’t  seem  to  share  my  bewilderment.  As  soon  as  she  called  for  Josh   and   Saheli   to   present   the   top   priorities   in   the   program   backlog,   I   waited   a  couple   of   moments   and   caught   Stephanie’s   eye,   quietly   pointing   at   the   door.   She  followed  me  out  while  Josh  began  his  emotional  speech  about  the  features  the  Train  will  take  on  in  the  next  ten  weeks.    “Yes?”  –  asked  Stephanie  in  her  typically  calm  voice.      “How  come  John  and  Tanya  left?  We  are  going  to  need  them.”    

Page 19: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   19  

“For  what?  I’m  sorry,  I  thought  we  needed  them  to  present  and  that  then  they  could  leave…”  She  realized  that  something  was  amiss.  She  sighed,  “Look,  I  told  them  it  was  ok.  Maybe  that  wasn’t  right.”    “It  wasn’t.  Their  thorough  participation  was  something  I  emphasized  when  we  were  discussing  the  agenda  last  week.  The  key  stakeholders  need  to  be  with  the  program  for  the  two  days  of  planning.  Josh  and  Saheli  are  managing  the  backlog,  but  they  are  not   the   ones   that   decide   in   principle  what   this   Train   is   building.  We   need   a   full-­‐fledged  business  owner  team,  which  in  this  case  consists  of  John,  Tanya,  Saheli  and  Josh.”    “I   understand   and   I’m   sorry.   Let’s   run   over   and   grab   them  and   see   if  we   can   talk  them   into   coming   back.”   She   sighed   again.   She  was   probably   thinking   about   how  difficult  it  would  be  to  convince  them  to  come  back,  especially  since  they  didn’t  like  the  idea  of  this  planning  in  the  first  place  and  now  here  she  was  asking  for  more…    If  that  was  her  thinking,  she  was  right  on  the  money.    “What?  Are  you  kidding  me?  I  can’t  spend  any  more  time  on  that,”  said  John  when  Stephanie   voiced   the   request.   “Tanya   and   I   spent   almost   an   hour   there   already.   I  don’t  see  any  reason  to  spend  more  time.”  Tanya  was  sitting  at  the  opposite  side  of  the  table,  nodding  in  agreement  as  John  spoke.      This   is   bad.   Without   these   two   key   people   the   program   will   drift   in   the   wrong  direction.  I  have  to  do  something  right  now  or  the  whole  thing  is  in  danger.  Teams  will  plan  something  that  these  two  guys  will  then  claim  is  wrong  in  ten  weeks  and  that  will   be   a   sad   fiasco   for   the  whole   initiative.  A   lot   of   the  Train’s   effort  will   be  wasted  as  a  result.      “John,  if  I  may…”  John  raised  his  eyebrows  and  stared  at  me.  “Look,  I  know  you  and  Tanya  must  be  busy  and  have  a  lot  of  things  going  on.  I  wouldn’t  expect  otherwise,  given   your   role   in   the   company.   But   I  wonder   if   you   noticed  what   just   happened  back  there  in  ‘Town  Hall’?”    John  looked  at  me  suspiciously.  It  is  important  that  I  give  him  the  facts  and  then  let  him   decide.   I’ll   do  what   I   can,   but   eventually   it   is   his   program,   not  mine,   so  who  should   be   more   worried   about   the   outcome   here?   “John,   your   speech   absolutely  riveted   their   attention.  This   could  be  a  pretty  noisy  audience  –   I   can   see   that.  But  there   was   no   chatter   or   fidgeting   or   side   conversation   going   on   while   you   were  speaking.  You  know  why?”    “Why?”        “Because   they   care…  There  were  many   things   that   you   and  Tanya   presented   that  they  heard  for  the  first  time  today.  That  was  an  important  piece  of  information  that  

Page 20: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   20  

they   needed   to   hear   from   you.   It   is   not   something   their   product   owners   will   tell  them.  Nor  will  they  figure  that  out  on  their  own.      “Things  that  are  intuitive  and  obvious  to  you  are  only  that  way  because  that  is  your  world.  You   think  and  operate   in   terms  of   the   strategic   intent   for   the  organization,  but  they  don’t.  They  are  stuck  in  their  specific,  narrow  boundaries.  But  guess  what?  Every  one  of   them   is  making   lots  of   little  decisions  every  day  and   those  decisions  add  up  to  either  good  or  not  so  good  outcomes.  When  a  developer  decides  how  to  round   a   numeric   value   of   the   variable,   or   what   response   time   is   sufficient   for   a  screen,   or  which   field   to   index   a   table   by,   there  needs   to  be   some  meaning,   some  context  for  those  decisions.  There  must  be  some  glue  to  make  the  parts  adhere  into  a  broader  vision,  for  a  successful  solution  that  will  benefit  the  consumer,  as  well  as  the  business.   If  a  developer  chooses   to  create  an   index  to  slightly  speed  up  search  queries   in   the   table,   and   that   table   turns   out   to   be   ‘write-­‐intensive’,   then   she   has  done  much  more  harm  than  good.  You  have  a  hundred  and  ten  people  in  there  that  need  guidance.  Imagine  how  many  wrong  decisions  they  will  make  within  the  next  ten  weeks  without  appropriate  context.  Imagine  how  much  waste  and  rework  there  will  be…”    John  sat  quietly  for  a  few  seconds  and  then  said,  “Look,  you  are  probably  right.  But  I  gave  them  that  just  a  few  minutes  ago.  I  gave  them  the  context  you  are  talking  about.  For  the  rest  of  today  and  tomorrow,  I  have  a  hundred  more  things  to  get  done  and  I  bet  so  does  Tanya…”    I   interjected:   “What   you   did   was   very   valuable.   But   it   was   a   one-­‐way   trip.   You  delivered  the  message  to  the  group  –  great!  But  that  doesn’t  mean  that  they  properly  understand  it  yet.  The  only  way  you  can  confirm  their  understanding  is  by  hearing  from  them;  when  you  see  their  plans,  their  PI  objectives.  They  need  to  process  your  message  and  provide  some  output  that  proves  whether  they  got  it  right  or  not.  And  trust   me,   based   on   prior   experience,   I   know   for   sure   that   you   will   find   some  significant  inconsistencies  with  your  initial  vision.  Would  you  choose  not  to  know?”    John   looked   at   Tanya,   a   little   puzzled.   He   opened   his   laptop   and,   it   seemed,  completely  tuned  us  out  for  a  minute  or  two.  He  finally  put  it  aside  and  looked  at  all  of   us  with   a   softened   expression.   “Very   busy   two   days,   guys.   I   can   probably   shift  some  things  around  but  don’t  expect  much.  I  can  give  you  about  an  hour  more  today  and  about  the  same  amount  of  time  tomorrow.”    That’s   far   from   perfect,   but   we’ll   have   to   live   with   it.   It’s   certainly   better   than  nothing.      “In  that  case,  John,  if  you  could  come  by  today  at  four  for  the  draft  plan  review  and  then  again  tomorrow  at  one  when  the  teams  present  their  final  plans,  that  would  be  perfect.”    John  nodded  and  looked  again  at  his  monitor.  “How  about  you,  Tanya?”  he  asked.    

Page 21: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   21  

 “I  think  I  can  make  those  time  slots  work  as  well,”  she  answered.    “Great!”  Stephanie  jumped  in.  “Thank  you  both  so  much  for  your  flexibility  on  this.  We  all  really  need  you  two  to  make  this  program  successful.”    The  door  closed  and  Stephanie  and  I  hurried  back  to  the  ‘Town  Hall’.      “Ryan,   I’m   sorry   for   this.   I   should   have   remembered   what   you   asked   for   earlier.  Honestly,  I   just  had  too  much  to  take  care  of  with  the  logistics  and  other  things  –  I  just  missed  it.”    “That’s  okay.  At  least  we  have  them  for  two  more  hours  now.  Let’s  make  sure  we  use  their  time  efficiently.  I  think  Josh  and  Saheli  must  have  finished  their  presentations  by  now.”    We  entered  Town  Hall  and  witnessed  just  the  opposite  –  Josh  and  Saheli  were  being  bombarded   with   questions   about   functionality,   nonfunctional   concerns,   and   so  forth.   The   teams,   realizing   that   they   finally   had   a   chance   to   resolve   the   questions  about  scope,  had  quickly  taken  advantage  of  the  opportunity.  After  four  or  five  more  questions,  Stephanie  was  able  to  usher  Saheli  and  Josh  away  from  the  podium  and  invite   the   architects   up.   This   session   went   full   length   as   teams   wondered   about  different   aspects   of   design   for   new   features,   tooling,   and   other   engineering  considerations.  With  just  fifteen  minutes  left  before  lunch,  I  took  over  the  meeting  to  cover  one  last  thing  before  the  team  breakout  starts.      “This  is  what  your  plan  is  going  to  look  like,”  I  pointed  at  the  wall  to  the  left  of  the  projector  screen.      

Page 22: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   22  

   “Every  team  will  have  seven  big  sheets.  Five  of  the  sheets  will  be  for  the  five  Sprints  in   the   PI.   This   is   where   you   will   put   your   stickies   –   the   stories   that   you   will  formulate,   once   you   start   breaking   down   features   into   smaller,  more  manageable  chunks.   Please   keep   in   mind   that   every   team   will   still   have   their   regular   Scrum  ceremonies  (including  Sprint  planning)  every  Sprint.  That’s  when  you  will  provide  an  even  deeper  level  of  detail  to  those  backlog  items.  But  for  now,  go  only  as  deep  as  you   need   to,   in   order   to   roughly   understand   the   size,   dependencies   and   overall  significant  risks.  In  fact,  as  I  said  before,  it  is  important  that  you  stay  at  a  high  level,  otherwise   you   will   definitely   get   stuck   in   the   details   and   never   manage   to   get  through  the  entire  PI  scope.  Each  sticky  will  have  a  story  size  estimate,   just   like  in  the  sample  stories  that  I’ve  created  here.  Each  Sprint  sheet  will  have  two  numbers:  Velocity   –   which   you   will   have   to   estimate   for   each   Sprint   in   the   PI,   taking   into  account  time  off  or  other  distractions;  and  Load  –  the  overall  amount  of  points  on  all  stories  loaded  into  that  Sprint.  Please  be  realistic  and  don’t  expect  miracles.  Those  two  numbers  are  designed  to  be  a  self-­‐check  for  each  team,  in  terms  of  reasonable  workload  planning.    “Backlog   items   will   be   color-­‐coded:   green   for   the   new   functionality,   orange   for  spikes   or   refactoring   effort,   purple   for   maintenance,   and   red   for   risks   and  dependencies.  Please  use  these  colors  consistently.    “Another  sheet  –  actually  the  most  important  one  –  is  the  list  of  your  PI  objectives.  No  stickies  on  this  one.  Instead,  each  team  will   literally  write  the  objectives  on  the  sheet  in  the  form  of  a  list,  clear  and  legible.    “The  seventh  and  final  sheet  will  be  for  risks  and  impediments  that  the  team  cannot  resolve  on  their  own.  These  need  to  be  brought  to  the  program’s  attention.  

Page 23: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   23  

 “We  will  adjourn  for  lunch  until  1  p.m.  But  it  would  be  helpful  if  you  guys  could  set  up   your   team   area   –   the   seven   sheets   –   so   that   we   will   have   three   full   hours   to  dedicate   to  planning.  Please   remember   that  your  goal   today   is   to  provide  a   rough  plan  of  all  five  Sprints,  enabling  you  to  derive  meaningful  PI  objectives  and  present  those  to  the  business  owners  at  four  o’clock.  Let’s  meet  here  at  one  sharp  for  a  quick  check-­‐in  and  to  kick-­‐off  the  actual  planning.”    Some  teams  went  directly   to   the  kitchen  area  –  clearly  processing   too  much   input  during   the   briefings   consumes   a   lot   of   energy.    Others   used   the   break   to   pick   the  best  parts  of   the  wall,   close   to   their   team   tables  and  smooth  enough   to  attach   the  sheets  without  worrying  about  switches,  thermostats  or  windows.  In  just  one  hour,  the   real   action   would   kick   off,   and   result   in   a   meaningful   plan   that   helps   them  realize,   through   experience,   that   they   are   really   one   big   team.   I   certainly  want   to  believe  that’s  the  case…          

Page 24: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   24  

IV.  BITTER  TRIUMPH        It   is  1:  15  p.m.,  and  most  of   the  teams  are  busily  circulating  around  their  planning  areas.  A   few   still   sit   at   their   tables  with  Product  Owners   trying   to   pull-­‐up   info  on  their   laptops.   Josh   and   Saheli   were   hanging   out   with   the   Avengers,   explaining  something   to   the   team.   Josh  waived   his   hands   in   his   typically   expressive  manner,  while  Saheli  made  notes  on  the  stickies  and  handed  them  to  the  team.  That’s  a  very  good  sign  –  that’s  why  they  were  brought  in  here  –  to  work  with  one  another.      The  whole  room  looks  and  sounds  like  a  busy  beehive.  Hopefully  the  product  will  be  equally  sweet  –.      “Ryan,   does   everything   look  good   so   far?”   Stephanie  had  approached   so   silently,   I  haven’t  even  noticed.      “Are  you  kidding?  They  are  only  fifteen  minutes  into  it,  so  how  can  I  tell?  However,  some  good  things  are  already  happening.  See  Saheli  and  Josh?”    “Yes…”    “Well,   they   are   exactly  where  we   need   them   to   be   during   the   breakout   session   –  helping  teams  and  addressing  their  inquiries.  It’s  too  early  to  get  real  excited,  but  in  the  meantime,  the  first  Scrum  Master  check-­‐in  is  going  to  happen  in  thirty  minutes  –  let’s  go  prep  the  area.”    It  is  key  for  the  entire  group  to  stay  focused,  and  in  order  to  stay  focused  you  need  a  few   key   process   goals   that   are   easy   to   follow.   Stephanie   eagerly   assisted   me   in  creating  those  and  now  it’s  almost  time  to  call   for  the  first  Scrum  Master  check-­‐in.  The  teams  are  so  busy  planning,  that  only  three  Scrum  Masters  come  without  being  prompted.  We  had  to  switch  on  the  microphone  again  to  call  for  the  rest  them.  Once  all  fourteen  of  them  had  gathered  around  the  check-­‐in  board,  we  get  started.      “Guys,  we  have   to  make   this   quick   and   efficient.  We  only  have   a   few  questions   to  check  status  on,  but   they  are   important  ones.   It’s  a  matrix  with  questions  as   rows  and  teams  as  columns.  I  will  pick  a  question  and  go  through  all  the  teams  and  then  another  question,  and  so  on.  All  teams  will  have  to  tell  me  where  they  are  and  then  we  will  move  to  the  next  question.  Once  we  complete  this  part,  hopefully  quickly,  we  will   have   a  meet-­‐after,   where   you   are   welcome   to   bring   up   specific   impediments  your  teams  have  discovered.    

Page 25: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   25  

   Let’s  get  started.  First  entry:  ‘Are  features  broken  down  into  stories?’  I’d  like  to  hear  from  you  guys,  team  by  team  now.  Avengers?”      “No,  not  quite  finished  yet,”  said  the  Scrum  Master  of  Avengers.  “We’re  still  working  on  it.”    “Good.  Partial  credit  then,”  I  said  and  drew  my  favorite  half-­‐dot  sign  in  their  cell  on  the  matrix.      “Next…  Argonauts?”    “Done!”    “Very  good.  Solid  dot  for  you  guys.”      “’Spartans’?”    “Not  done,  but  close.”    We  are  moving  really   fast  with  this.  The  rest  of   the  teams’  Scrum  Masters  provide  their  progress  quickly  and  we  move  to  the  next  question:      “’How  many  Sprints  planned?’  Guys,  I  need  to  warn  you  to  be  careful  with  this  one.  ‘Planned’  means  a  very  specific  thing  for  the  Sprint.  It  means  it  has  stories  loaded  on  the  Sprint  sheet  and  it  has  two  numbers  on  it  –  velocity  and  load.  Then  it  counts.  I’m  not  trying  to  be  overly  formal  here,  but  with  fourteen  teams  in  the  room  we  need  to  be  disciplined   about   the  process   to   avoid  unproductive   chaos.   I  will   be  putting   as  many  dots  in  a  cell  as  the  number  of  Sprints  you  have  planned.  Once  you  have  five  dots  –  it’s  going  to  be  a  check  mark…  simple!  So,  Avengers?”  

Page 26: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   26  

 “None  yet.”    “Okay.  Next.  Argonauts?”    “One   done,”   their   Scrum   Master   proudly   pointed   to   their   team   area.   It   had  everything  in  place,  including  velocity  and  load  for  the  first  Sprint.      “Good  job,  guys.  One  dot  out  of  the  way…”    The  rest  of   the  process  went  quickly.  Partly  because  not   too  many  Scrum  Masters  had  much  to  share.  This  is  a  good  exercise  to  show  them  where  they  are,  relative  to  the  goal  for  the  day  –  to  have  a  plan  ready  to  present  to  the  business  owners.      

   “Guys,   please   look   carefully   at   our   check-­‐in   sheet.   Your   big   next   goal   is   to   finish  loading   scope   into   the   Sprints   and   formulate   your   draft   PI   objectives.  Meanwhile,  make   sure   you   start   effectively   identifying   risks,   impediments,   and   dependencies  with  other  teams.  That’s  it  for  the  main  part.  We  finished  in  nine  minutes  –  not  bad  at  all  for  the  first  time.  Next,  is  the  meet-­‐after  if  we  need  one.  Are  their  any  problems  you  see  so  far?”    “Yes,”  said  the  Avengers’  Scrum  Master.   “Just  before  coming  here  my  guys  asked  a  very  good  question  that  made  me  rethink  the  way  we  are  planning  the  PI.”    “Who  do  we  need  for  this  discussion,  do  you  think?”    “I’m  afraid  we’ll  need  all  Scrum  Masters  because  it  is  going  to  affect  the  entire  Train.  I  also  need  Sunil  –  a  developer  from  my  team…”    

Page 27: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   27  

Stephanie  quickly  went  in  search  of  Sunil,  to  bring  him  into  the  conversation.  In  the  meantime,  the  Scrum  Master  went  on:    “We  actually  almost  finished  planning  Sprint  one  and  had  started  working  on  Sprint  two,   when   Sunil   told   us   that   most   of   the   stories,   he   thinks,   are   considerably  underestimated.  Here  he  is  actually,  so  I’ll  let  him  speak…”    “All   the  stories  we  have  on  those  sheets,  except   for  spikes,  will   take  much  longer,”  Sunil  quickly   jumped   in.   “I  realized  that  we  did  not  account   for  end-­‐to-­‐end  system  integration.  During  the  last  few  weeks,  we  were  experimenting  with  it  because  Ryan  raised   the   issue  of   full   system   integration  and  we  had  agreed   to   integrate  at   least  three   times  per  Sprint.  Unfortunately,  we  never  bothered   to  determine  how  much  more  time  it   takes.  But  we  know  how  painful   it  was  during  our   initial  experiment.  The   other   guys   and   I   believe   that   it  will   take   ten   to   thirty   percent   for   each   story,  depending  on  the  complexity  and  the  number  of  teams  involved.”    Some  of  the  Scrum  Masters  nodded  in  agreement.  Stephanie  glanced  at  me  and  then  proposed  an  idea:  “Why  don’t  we  all  simply  assume  an  average  of  20%  to  be  added  to   the   current  estimates.  Every   team  will  have  dozens  of  backlog   items   for   this  PI  and  I  think  the  law  of  large  numbers  will  make  it  work.”    She’s  right.  We  need  to  communicate  this  to  the  teams  ASAP.      “We  will  need  every  Scrum  Master  to  communicate  this  to  their  team.  The  message  is  simple  –  account  for  roughly  20%  more  to  cover  integration  effort  for  each  story,  except   for   spikes,   other   research   or   anything   that   does   not   need   to   be   integrated  into  the  mainline.  Are  we  clear  on  this  guys?”    All  the  Scrum  Masters  agreed  and  now  we’re  ready  for  action  again.      “Wait!”  Another  Scrum  Master  raised  his  hand.  “My  team  has  a  serious  problem.  It  affects  only  us  and  Spartans  and  we  had  better  bring  in  architects  for  this  one.”    Stephanie   rushed   to   get   Bill   and   Todd,   the   system   architects   working   with   this  group,  who  were  sitting  idle  at  their  table  in  the  back  of  the  room.        “Guys,  why  don’t  you  all  quickly  go  back   to  your   teams  and  notify   them  about   the  change  in  the  estimation  of  stories.  I  will  need  Ninjas’  and  Spartans’  Scrum  Masters  back   here   immediately   afterward   for   the   session  with   the   architects.   Feel   free   to  bring  other  team  members  as  you  see  fit.”    In  less  than  five  minutes,  the  entire  group  of  seven  people  was  discussing  the  issue  that   Ninjas   had   uncovered.   It   turned   out   that   this   was   the   first   time   they   had   to  apply   a   series  of   changes   to   the  database   structure   that  would  affect   a  number  of  tables   with   high   transaction   intensity.   The   tables   are   basically   responsible   for  storing  data  for  the  major  part  of  the  purchasing  process.  Data  transformation  and  

Page 28: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   28  

transfer   to   the  new   structure  was   estimated   to   take   at   least   thirty  minutes   in   the  production   environment.  With   an   average   rate   of   three   hundred   transactions   per  second,   they   needed   a   smart   way   to   handle   this.   Luckily,   Todd   was   at   hand   and  suggested  creating  a  temporary  table  to  capture  all  the  transactions  while  the  main  data  transfer  process  was  running  in  the  background.  The  maintenance  window  can  be  then  reduced  to  less  than  a  minute  –  just  enough  time  to  flush  all  the  data  from  the  temporary  transaction  tables   into  the  main  tables,  once  the  main  data  transfer  process  is  complete.      “We  are  looking  at  three  minutes  though,  including  the  time  to  restore  indices.”    That  was  a  good  catch  by  Ninjas.  Now  they  will  have  to  add  additional  items  to  the  plan   to   create   data   transformation   scripts   that   would   automatically   handle   the  process,   and   test   it   very   thoroughly   for   any   errors   in   transaction   processing   that  could  be  fatal  to  the  business.      The  architecture  session  ended   there.  Stephanie  and   I   finally  got  a   few  minutes   to  silently   observe   the   movement   in   the   room   without   arranging   any   sessions   or  helping  teams  exchange  thoughts,  or  what  not.  It  appears  that  after  the  initial  check-­‐in,  there  is  more  and  more  action  in  every  team  area.  Josh  and  Saheli  split  up  and  are  working   with   different   teams,   busily   clarifying   and  moving   things   around   on   the  sheets.  Both  Bill  and  Todd  were  helping  Spartans  with  estimation.      I  could  sit  and  watch  something  like  this  forever,  but  there  were  other  things  to  do.  The  next  check-­‐ins  introduced  more  interesting  questions,  but  teams  seemed  to  be  able   to  move   forward.   Scoping   decisions   after   the   second   check-­‐in   unblocked   the  Samurais,   who   otherwise   could   not   fit   anything   else   in   the   PI.   A   few   more  implementation   nuances   required   Bill’s   attention,   but   everything   was   resolved,  since  Bill  had  moved  to  helping  another  team.      The   last   check-­‐in   looked   really   encouraging.   All   teams   managed   to   get   to   the   PI  objectives.  And  while   there  were   small   issues  with   some   teams,   it   looked   like   the  Train  was  ready  to  present  the  plan.      

Page 29: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   29  

   John   and   Tanya   showed   up   a   few  minutes   before   four   o’clock,   both  with   curious  expressions   as   fourteen   teams   came   into   view   in   front   of   them.   All   were   busily  discussing   scope,   walking   around   to   other   teams,   and   finalizing   the   plans   for   the  initial   review.  The  walls   looked   like  some  sort  of  mosaic,   tiled  out  of  stickies.  Red,  orange,   purple,   green   –   they   all   harmonized   into   a   peculiar   ornament   that   was  nearly  ready  to  tell  the  story  of  Pacific  Express.  For  someone  who  had  left  the  room  in   the   morning   when   the   walls   were   entirely   empty,   this   picture   would   indeed  appear  quite  shocking.      Now   it’s   time   to  wrap  up  and  present  plans.  Most   teams  decided   that   the  Product  Owners  would   be   the   ones   presenting.   Stephanie   shared   the   presentation   agenda  with  the  group.  Each  team  will  present  the  PI  objectives  and  the  overall  flow  of  the  plan,   including   key   impediments   and   dependencies   if   any   have   been   identified   so  far.      The   Spartans  will   present   first.   John,   Tanya,   Saheli   and   Josh   stand   as   close   to   the  Spartans  team  area  as  they  can  in  order  to  follow  the  outline  of  their  PI  scope.      “In  this  PI,  the  team  is  going  to  pursue  the  following  objectives:    

-­‐ Coupon  entry  at  Point-­‐of-­‐Sale  -­‐ Mirrored  real-­‐time  transaction  data  transfer  scripts  -­‐ Support  of  different  coupon  types  -­‐ Research  and  prototyping  of  earned  points  algorithm  -­‐ Basic  discount  points  for  registered  customers  

 “Our  estimated  velocity   in  Sprint  1   is  45  and  the   load  is  42.  Velocity   in  the  second  Sprint  is  40  with  the  load  of  39  story  points…”      

Page 30: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   30  

 

   He   spent   another  minute   describing   capacity  matching   and   giving   the   audience   a  rough  idea  of  what  gets  done  in  each  Sprint.      “We  have  dependency  on  the  Data  Warehousing  team,”  he  said,  finishing  his  talk.      Stephanie  brought  the  second  mic  and  stood  closer  to  the  business  owners.    “I  have  a  question,”  said  John.  “What  is  that  Mirrored  data  transfer  thing?”    “Basically,   in   order   to   be   able   to   apply   a   coupon   at   the   point   of   sale,  we   need   to  change  the  database  structure,  where  affected  tables  have  very  high  transaction  rate  per  second.  This  requires  us  to  use  mirroring  tricks  to  capture  all  transactions  that  will  occur  in  the  system,  while  the  main  corpus  of  data  is  being  reloaded  into  a  new  structure  of  tables.”    “Okay,”   John   said   without   too   much   confidence.   “Are   you   sure   this   has   business  value  to  the  organization?  I  definitely  agree  that  it  is  a  necessary  step  to  manage  the  data  transfer  once  the  functionality  is  ready,  but  is  it  useful  enough  to  be  called  an  objective?”    Stephanie   quickly   jumped   into   the   conversation:   “John,   let   me   describe   what’s  happening  here.  What  he  just  explained  is  not  a  one-­‐time  process.  They  are  creating  a  set  of  scripts  that,  once  developed,  can  be  used  repeatedly  for  purposes  like  this.  Also,  the  scripts  can  be  used  with  other  tables  with  a  minimum  of  modification.  So  

Page 31: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   31  

we   are   really   talking   about   a   re-­‐usable   capability   that   will   recurrently   allow   our  company  to   implement  the  new  functionality  and  always  have  a  way  to  reload  the  data  within  an  extremely  tiny  maintenance  window.”    John  nodded.  “I  see…  I  actually  agree  that  it  has  value.  Thanks  Stephanie.”    “Any   questions   from   the   teams?”   I   turned   to   the   rest   of   the   tables   where   team  members   were   all   sitting   together,   just   like   they   had   done   in   the   morning.   After  three   hours   of   highly   intensive   planning   effort,   they   are   all   curious   to   see   what  others  have  to  present.      “I  have  a  question,”  shouted  a  guy  in  a  yellow  t-­‐shirt  with  code  snippets  on  it.  “What  kind  of  dependency  do  you  have  on  us?”    As   the   Spartans’   Product   Owner   began   to   explain,   it   became   apparent   that   the  dependency  had  not  been  communicated   to   the  other   team.  Stephanie  didn’t  miss  this  opportunity  to  remind  everyone  that  they  are  supposed  to  not  only  identify  the  dependencies,  but  also   to  resolve   them.  Nevertheless,   the   team  definitely  deserved  applause  and  the  entire  room  exploded  in  cheers.  We  are  moving  to  the  next  team.      The  next  presentation  seems  to  go  quite  well  and  John  does  not  have  any  questions.  Saheli   ponders   aloud   some   details   of   the   functionality,   but   that   was   it   from   the  business  owners’  side.  Surprisingly  though,  this  team  also  mentioned  a  dependency  on  data  warehousing   team.  Unlike   the  previous  case,   it  was  communicated   to   that  team,  but  it  was  also  much  bigger  and  riskier  in  nature.      Team  after  team,  we  gradually  managed  to  go  through  the  entire  group.  Some  good,  some  not  so  good,  but  every  team  had  their  PI  objectives.  Some  teams  ended  up  with  an   unrealistic   ratio   of   velocity   and   load,   which   was   pointed   out   and   taken   as   an  action  item.  The  story  with  data  warehousing  did  not  end  up  on  the  first  two  teams.  With  growing  focus  on  retail  analytics  as  a  strategic  theme,  the  Train  ended  up  with  nine  teams  requiring  assistance  of  the  data  warehouse  team,  mostly  for  logging  the  data   that   newly   implemented   user   scenarios   would   imply.   This   substantial  dependency  issue  made  Stephanie  very  uncomfortable  with  the  plan.      In  the  meantime,  John  looked  very  puzzled.  On  one  hand  he  had  the  opportunity  for  the  first  time  in  his  career  to  see  what  teams  are  really  intended  to  do.  On  the  other  hand,  something  was  deeply  troubling  him  about  what  he  just  seen.  A  few  moments  later  he  grabbed  the  mic:      “Guys,  I  have  a  serious  concern  here.  I  realized  that  we  are  missing  a  lot  in  terms  of  the  pre-­‐order  functionality.  This  is  a  strategic  direction  for  us.  Some  chains  have  this  functionality  already,  while  we  haven’t  been  able  to  deliver  much  of  anything  lately.  To  compete  with  the  rest  in  the  race,  we  need  to  make  a  really  bold  statement  that  will   differentiate  us   and  make  our   customers   eager   to  use   that   functionality,   over  

Page 32: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   32  

and   over   again.   We   need   to   give   them   control   over   the   process   and   we   need   to  enhance  the  transparency  for  these  deceptively  obvious  user  scenarios.”    John  is  apparently  eager  to  discuss  this  further,  but  that’s  what  the  next  session  is  for.  We  need  to  let  the  teams  go  for  the  day.      “You  guys  did  a  great  job  today,”  said  Stephanie  smiling,  but  obviously  tired.  She  did  a  good  job  as  well  today,  that’s  for  sure.  Whether  or  not  she  will  be  a  facilitator  for  Pacific  Express  –  I  can’t  say,  but  she  definitely  understands  this  process  and  is  capable  of  pushing  things  forward.  “Sleep  well  and  see  you  all  at  eight  tomorrow,”  she  said,  adding:  “We  are  still  going  to  need  Scrum  Masters  and  Product  Owners  for  another  hour  today  for  the  Management  Review  and  Problem  Solving  session.”    While  the  teams  packed  their  backpacks  and  left,  the  rest  of  us  swarmed  around  the  whiteboard.  I  explained  the  process  to  the  group  that  is  staying  for  this  meeting:    “We  need  to  go  though  all  outstanding  issues  and  come  up  with  the  suggestions  that  we  will   communicate   to   the   teams   first   thing   tomorrow  morning.   Let’s   list   all   the  topics…”    “Pre-­‐order  functionality.  There’s  such  a  gap  here  that  I  can’t  believe  it!”  said  John  in  a   very   emotional   tone.   “I’m   also   changing  my   schedule   and   canceling  most   of  my  meetings  for  tomorrow.  I  don’t  want  to  come  just  to  the  final  presentation  and  find  out  that  the  teams  planned  something  way  out  of  alignment  with  the  key  themes.  I  want  to  be  here  during  the  process  itself,  not  only  for  the  final  review!”    Stephanie  gave  me  a  funny  look.  This  is  a  perfect  development.  John  realizes  that  the  teams   need   him.   He   probably   does   not   remember   what   happened   less   than   five  hours   ago   in   his   office,   when   he   explained   to   us   how   busy   a   person   he   is   in   this  company…  Ultimately,  who  cares?  What  matters  is  that  he’s  here  now,  and  he  wants  to  have  even  more  interaction  with  the  teams.      Another   topic   on   the   list   was   multiple   dependencies   on   data   warehousing.   John,  however,  quickly  got  us  back  to  the  previous  topic:      “We  need  them  to  add  more  pre-­‐order  functionality.  I  want  to  see  options  for  user  pick-­‐up  and  delivery,  as  well  as   for  curb  pick-­‐up.   I  also  want  order  status  tracking  and  status  updates  over  SMS.  We  need  to  show  our  customers  that  this  is  a  fast  and  transparent  process  with  multiple  options.  I  also  want  automatic  selection  of  closest  store   location,  and  selection  of  an  arbitrary   store  on   the  map.  We  need   to  make  a  strong  impression,  guys.”    Saheli  and  Josh  promised  to  work  with  the  teams  the  next  day  to  add  all  of  this  to  the  plan.  “It  means  that  we  will  have  to  pull  something  out  of  the  plan,”  said  Saheli.  “But  since  you  are  going  to  be  here  most  of  tomorrow,  we  can  probably  choose  the  items  to  de-­‐scope  together.”  

Page 33: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   33  

 John   agreed   to   that   plan   of   action   and   we   moved   to   the   second   topic   –   data  warehousing.      “I  don’t  think  any  of  this  will  execute  well  during  the  PI  if  one  team  has  almost  every  other  team  on  the  Train  depending  on  them,”  said  Stephanie,   looking  to  the  others  for  more  opinions.    “Guys,”  I  had  to  help  move  this  topic  in  the  right  direction,  “It  is  quite  obvious  that  the  Data  Warehousing  team  is  going  to  end  up  with  a  pretty  random  backlog  across  all  the  Sprints  with  that  many  dependencies  in  flight.  It  is  impossible  to  realistically  plan  and  even  worse  –  to  execute.  This  is  not  going  to  fly,  I’m  afraid…”    “Maybe  we  need  to  make  sure  we  invite  other  Scrum  Masters  to  this  team’s  Sprint  planning  every  time…  just  to  make  sure  they  are  synchronously  handling  it,”  offered  Saheli.    “Maybe  we  just  need  to  split  them  up  –  like  we  did  with  the  maintenance  team  –  and  have  those  people  join  the  teams  that  have  the  highest  number  of  dependencies  on  data  warehousing,”   said   Josh  as  he   approached   the  board.   “How  many  developers  are  on  the  team  now?”    “Six,”  said  their  Scrum  Master.  “And  one  tester…”    “See?”  Josh  smiled  widely.  “They  don’t  even  have  enough  testers  to  make  sure  they  are   doing   the   right   thing.   If   we’d   move   those   six   guys   to   other   teams,   their  functionality   could   at   least   be   tested   in   the   context   of   other   important   scenarios  those   teams  are  working  on.  Well,   this   is   the   smartest   thing   I’ve  heard   today!”  he  concluded,  breaking  into  laughter,  along  with  the  rest  of  the  group.  “Also,  they  didn’t  even   have   a   proud   name,   like   Spartans   or   some   other   team   of   heroes…   What   a  shame.”    People   are   still   chuckling,   but   Josh   obviously   has   a   point   even   though   it   was  delivered   in   his   typically   pompous  manner.  As   Stephanie  had  noted   in   one  of   our  private  conversations,   Josh   is  usually   in  one  of   the  three  states:  sarcastic,  cocky  or  simultaneously  both.  But  regardless  of  that,  he  really  gets  it  and  this  time  he  again  advised   the  correct   course.    As   it   turned  out,   the   team  didn’t  even  have  a  Product  Owner,  and  the  Scrum  Master  can  be  efficiently  used  with  the  newly  formed  System  Team  that  needs  one  very  badly,  but  doesn’t  have  one.  The  decision  was  supported  by  everyone  and  would  be  announced  the  next  day.    

*  *  *    An  hour  later  I  was  on  the  trail,  running  my  5k  down  the  beautiful  shores  of  Del  Mar.  The  ocean  looked  like  a  huge,  live  mirror,  breathing  and  spitting  out  splashes  of  the  pure,   late   evening   tides.  My  brain  was   slowly   getting  back   to  normal   and  my  GPS  

Page 34: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   34  

watch  was  still  giving  me  hope  that  I  could  cover  the  distance  in  twenty-­‐five  minutes  today.   My   running   belt   pouch   rhythmically   scratched   my   back   in   time   with   my  pace…    Wait…  it  is  actually  my  phone  vibrating.  Someone  is  calling.  I  quickly  rotate  the  belt  180  degrees  moving  the  pouch  to  the  front  –  I’m  trying  to  grab  my  phone  without  slowing  my  pace.  It’s  Stephanie…  Oh…  Not  now!  Now  I  have  to  stop…    “Ryan,   we   have   a   problem.   Do   you   remember   some   guys,   the   development  managers,  that  you  met  quite  a  while  time  ago  at  one  of  our  first  discussions?”    “Yes,  I  think  so…”  my  brain  busily  scans  for  a  picture.      “Well,  I  wonder  if  you  noticed  them  in  ‘Town  Hall’  today?”    “No,  I  don’t  think  so…”    “Turns  out  they  were  there.”    “But  wait   I  don’t  remember  them  at  our   last  session.  We  only  had   like  twenty   five  people  total  –  I  would  have  noticed  them.”    “See,   that’s   the   problem,”   Stephanie’s   voice   sounded   even   more   worried.   “They  stayed   during   the   breakout   session   and   then   for   the   plan   review,   but   left   right  afterwards.  In  fact,  they  did  not  leave  the  office.  Do  you  remember  seeing  Tanya  at  the  problem  solving  session?”    “No.  I  remember  her  and  John  during  draft  plan  review,  but  she  wasn’t  participating  much.”    “Right.   So   after   that   session   she   and   her   development  managers   conspired   in   her  office,  complaining  about  this  whole  initiative  and  the  planning  process.”    “What?!”    “Yes.  They  complained  to  her  that  with  the  new  process,  they  feel  like  they  are  not  needed  anymore.  With  teams  planning  on  their  own,   they   feel  nobody  cares  about  their  opinion.”    “Crap…”    “Yep.  Tanya  called  me  and  voiced  the  seriousness  of  her  concerns,  saying  that  she  is  going  to  talk  to  our  CIO  and  express  them  to  him.  We  need  to  do  something  ASAP.  Ryan,  you  and  I  need  to  meet  at  the  office  tomorrow  at  seven  and  think  through  the  next  steps.  Can  you  do  that  for  me  please?”    

Page 35: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   35  

She  sounded  totally  desperate.  But   it  was  depressing   indeed.  We  had  had  a  pretty  good  day,  but  with  an  unexpected,  bitter  ending.      “Listen,   Stephanie,   I   have   an   idea.  We  will   need   to   talk   to   all   of   them   tomorrow,  including  Tanya.  I  have  something  to  present  to  them.  I’m  sure  we  have  a  chance  to  fix  this  misunderstanding.”    “We  better   have   a   chance,”   she   said   in   a   very  depressed   voice.   “Thank   you,  Ryan.  Thanks  for  helping  with  this.  I’m  really  worried  about  it  all.  Thank  you  and  let’s  do  all  we  can  tomorrow,  okay?  Thank  you.”    She  hung  up,  but  her  deeply  disturbed  voice  still  echoed  in  my  mind.  This  is  a  very  unnecessary  complication,  but  we  will  give   it  our  best  shot.  We  will  delve   into  the  heart  of  the  problem,  instead  of  running  from  it.  Tomorrow  she  and  I  will  win  big…  or  lose  big,  but  tonight  I  don’t  even  want  to  think  about  it.                

Page 36: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   36  

V.  TROJAN  HORSE    “I  feel  like  I’m  sending  a  sheep  down  among  the  wolves,  Ryan.  I’m  sorry.  But,  you  are  right,  the  Train  needs  me  in  Town  Hall.”    “Stephanie,  have  you  ever  heard  of  a  sheep  that  can  bite?  A  sheep  that  hunts  to  kill?”      “Very   funny.   I   talked  with  Tanya   this  morning.  She  agreed   to  at   least   listen   to  you  first,  and  then  decide  what  to  do.  That  means  that  she’s  not  going  to  go  to  the  CIO  and  bitch  about  us  right  away,  she’ll  give  it  a  chance…  she  probably  wants  to  savor  it…”  Stephanie  smiled,  but  the  smile  was  pretty  pathetic.      “Don’t  worry  about  me.  I’ll  do  what  I  have  to  do.  Please  make  sure  the  changes  to  the  team  composition  and  scope  of  work  are  properly  communicated  before  you  kick  off  the  breakout  session  today.  And  remember,  they  have  to  get  to  the  stretch  objectives  and  John  must  assign  business  value  to  all  PI  objectives  the  teams  have.”    Stephanie   nodded   and   smiled  without   saying   a  word.  Only   now  did   I   realize   how  different   our   positions   are   and   what’s   at   stake   for   her.   I   certainly   want   this  implementation   to  be  successful  and  will  do  all   I   can   to  make   it  happen,  but   she’s  probably  putting  her  whole  career  at  stake  here.  What’s  the  worst  that  can  possibly  happen  to  a  consultant  whose  method  is  not  adopted?  They  never  invite  me  again?  Big  deal.  But  Stephanie  could  easily  lose  her  job  in  a  political  fight  like  this.  I  can’t  let  that  happen…    “Guys,  you  all  know  Ryan.  He’s  here  with  us  today  to  address  our  current  concerns  with   the   process.   Feel   free   to   ask   questions   –   that’s   why   we   set   up   this   session.  Right,  Ryan?”    “That’s   right.   Thank   you,   Tanya.   Actually,   before   we   get   started  with   questions,   I  want  to  present  some  initial  thoughts  to  you  guys.  Some  things  to  think  about,  some  things  to  chew  on.”      Tanya  and  her  crew  were  sitting  at  the  same  table,  their  body  language  saying  more  than   a   thousand   words   possibly   could.   Some   had   their   arms   folded   across   their  chests,  others  were  settled  deep  in  their  chairs  –  all  of  them  sported  pretty  annoyed  expressions.  Awesome…  Should  be  piece  of  cake  to  melt  this  iceberg.  I  can  admit  it  to  myself  now  –  I’m  really,  really  worried.  But  I  move  forward:      “We  are  here  because  of  a  fundamental  question:  what  is  the  role  of  the  leader  in  a  Lean-­‐Agile  enterprise?  Many  of  you  have  operated  in  an  environment  that  employed  some  team-­‐level  Agility.  All  teams  were  using  Scrum  as  their  basic  process.  But  do  you  know  what  Scrum  says  about  management?”    

Page 37: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   37  

The  awkward  silence  drives  me  crazy.  I  need  them  to  participate,  instead  of  showing  how  disconnected  they  are.      “Guys?..”    A  few  more  moments  pass  in  complete  silence.    “Nothing?”  The  guy  sitting  closest  to  me  guesses.  “It  says  nothing?”    Oh,  thank  goodness  for  any  answer  at  this  point,  because  the  right  answers  up-­‐front  is   not   what   matters   in   our   little   Socratic   experiment   here.   I   just   need   them   to  abandon  all  bias  and  start  thinking.    “Close.  But   it’s  actually  a   little  worse  than  that.  Have  any  of  you  heard  the  story  of  the  Chicken  and  the  Pig?”    One  person  nodded:  “It  says  we’re  chickens,  because  we  are   ‘not  committed  to  the  successful  outcome’  for  the  team.  At  least  it  doesn’t  call  us  pigs.”    The  others  chuckled.    “You   have   probably   also   heard   that   Scrum   Master   is   the   one   who   removes   all  impediments  for  the  team…”    They  all  nodded.      “So,  how  many  real  impediments  in  this  program  did  they  remove  so  far?”    Again,  they  chuckled.      “Unfortunately,  at  scale  this  turns  into  a  rhetorical  question.  But  please  don’t  think  that   this   is   because   your   Scrum  Masters   are   bad,   or   Scrum   per   se   is   misleading.  Scrum   is  a  great  method   that  undoubtedly   revolutionized   the  way   teams  work.  At  scale,   however,   different   forces   come   into   play.   For   that   reason,   leadership   is   a  whole  separate  domain  in  Scaled  Agile  Framework.  It  provides  critical  guidance  to  those   who   actually   can   eliminate   systemic   impediments   in   large   programs   like  Pacific   Express   –  managers   that   have   sufficient   authority   to   change   things   for   the  better.”    The  posture  and  the  expressions  on  their  faces  began  to  change  from  challenging  to  confused.  I  proceeded:    “Teams  always  try  to  do  their  best.  But  they  are  the  hostages  to  the  system  they  are  part  of.  We  know  this   from  our  own  decades  of  experience.   If  you  put  people   into  functional   silos,   no   matter   how   hard   they   work   the   value   will   flow   very   slowly  through   the   system,   and   the   resulting   quality   will   be   unacceptable.   Do   you   think  

Page 38: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   38  

those  departments  could  themselves  re-­‐organize  into  cross-­‐functional  teams  in  any  of  those  organizations?  Did  it  happen  on  its  own  in  yours?”    Now   it   looks   like   they  have  started  paying  even  more  attention.   I  have   to  proceed  down  this  path.  This  problem  has  to  crack.      “I  doubt  it.  I’m  sure  management  made  the  decision  and  assisted  in  identifying  the  new  team  composition  for  Agile  teams…”    The  guy  right  next  to  me  –  the  one  so  familiar  with  the  Scrum  folklore  –  nodded  and  looked  around  the  room:  “He’s  right.  We  actually  did  the  transformation,  although  it  only   affected   a   few   teams.  The   rest   of   the  program  was  hired   afterwards,   and  we  already  knew  how  to  ‘package’  them  into  Scrum  teams.”    “Good.  Now,  are  you  aware  of  the  concerns  product  people  have  with  respect  to  the  program?”    “We  are,”  he  nodded  again.      “Would   you   agree   that   given   current   structure   and   process,   the   teams   are   ideally  suited   to   continue   producing   their   own   chunks   of   value   independently   and,  conversely,  not  driven  at  all  to  synchronize  and  integrate  across  the  board.”    “Yes…”    “If  you  folks  will  not  support  them  in  the  adoption  of  practices  that  foster  alignment  and   incremental  development  of   the  whole   solution,   they  will   fail,   but   through  no  fault  of  their  own.  Stephanie  said  that  you  were  concerned  that  teams  are  planning  without  you;  that  you  have  nothing  to  do  now  with  the  advent  of  the  new  process?”    The   looks  on   their   faces   shifted   from  puzzlement   to   shock.  We  will   either   resolve  this   once   and   for   all,   or   I   will   be   thrown   out   of   this   room.   Tanya   remained   very  suspicious.      “Last  night  I  prepared  some  bullet  points  for  you  guys,”  I  hooked  up  the  HDMI  cable  to  my  laptop  and  the  title  slide  gradually  showed  up  on  the  projector  screen:  ‘What  Do  I  Do  as  Lean-­‐Agile  Leader’.  I  hit  the  clicker  and  moved  to  the  first  slide:    

• Help  the  teams  advance  Scrum  practices  • Help  with   Scalable   engineering   practices,   namely:   frequent   integration   and  

automation  • Help  remove  organizational  impediments  that  impede  the  flow  of  value  • Teach  the  teams  vertically  slicing  business  value  • Engage  the  business  and  upper  management  in  active  alignment  with  teams  

 I  move  to  the  next  slide:  

Page 39: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   39  

 • Understand  role  bottlenecks  and  provide  better  balance  of  force  • Help   teams   eliminate   long   term   dependencies   through   cross-­‐discipline  

orientation  • Lead/establish/support  communities  of  practice  • Help  the  program  in  visualizing  WIP,  variability,  and  hidden  dependencies  • Coach  teams  in  effective  continuous  improvement  techniques  

 …and  the  next  one:    

• Help   upper   management   and   business   understand   and   embrace   Lean  thinking    

• Understand  and  exploit  specific  nuances  of  knowledge  creation  and  sharing  in  the  program  

• Identify  key  factors  that  build  trust  between  the  teams  and  the  business  • Create  a  tangible  improvement  roadmap  for  the  program  and  lead  the  teams  

through  the  improvement  journey  • Foster  team  and  program  spirit  of  self-­‐organization  

 The  expressions  on  their  faces  began  to  change.  “So,  does  anybody  still  think  there  is  a  deficit  of  involvement  for  a  leader  in  a  Lean-­‐Agile  enterprise?  Oh,  but  wait,  there  is  more!”  I  hit  the  clicker  again:    

• Help  teams  engage  into  highly  collaborative  intra-­‐  and  inter-­‐team  workshops  in  identifying  requirements,  elaborating  design,  etc.  

• Provide  necessary  guidance  on  soft  skills  and  collaborative  techniques  • Enable  spontaneous  situational  leadership  among  the  team  members  • Improve/influence/change   ‘outside’   (with   respect   to   the   teams)   processes  

such  as  deployment,  compliance  validation  etc.  Shorten  external  components  of  the  overall  lead  time  in  end-­‐to-­‐end  value  delivery  

• Create   a   balanced,   decentralized   decision   making   model   that   covers   a  majority  of  the  risky  areas  

 …And  again:    

• Engage   all   necessary   stakeholders   in   program   backlog   refinement   and  economic  prioritization  

• Foster   collaborative   culture   within   the   teams   (side-­‐by-­‐side   work,   rotation  over  functionality,  etc.)  

• Keep   external   teams   and   individuals   aligned,   motivated   and   capable   of  collaborating  with  the  program  

• Assist  programs  in  exchanging  practical  knowledge      “And  certainly  there  are  many  more  real   life  examples  that  I  did  not   include.  After  Stephanie’s  call  yesterday,  I  timeboxed  myself  to  twenty  minutes  and  came  up  with  these  five  dense  slides.”  

Page 40: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   40  

 The   gentleman   next   to   me   rubbed   his   forehead   and   said:   “That’s   a   lot   of   things  indeed,  Ryan.”    “It  is  a  lot,  you’re  right.  But  let  me  tell  you  the  trick  we  use  to  get  that  many  touch  points   in   a   Lean-­‐Agile   environment.   Did   you   notice   a   single   item   on   that   list   that  would  actually  have  anything  to  do  with  managing  the  teams  per  se?”    “No…”    “Exactly.  The  principal  force  behind  these  dozens  of  bullet  points  is  the  switch  from  managing   to   enabling   teams.   And   since   enabling   teams   is   what   we   are   after,   we  naturally  arrive  at  a  gazillion  aspects  of  the  Leader  role.  As  you  shift  your  thinking  and   start   looking   for   the   bottlenecks   that   prevent   the   Train   from   fast   delivery   of  value  to  the  business,  you  arrive  at  more  and  more  things   like  that.  Take  a   look  at  these   again…”   I   browsed   through   the   presentation   once   again,   pausing   a   few  seconds  on  each  slide.  “Do  you  think  teams  can  do  any  of  these  on  their  own?”    “No,  Ryan,  I  understand  your  point,”  the  same  gentleman  said.  “I  don’t  know  about  you  guys,”  he  turned  to  the  rest  of  the  group,  “but  I  think  we  are  in  the  wrong  room  right   now.  Our   teams  need  us   in   ‘Town  Hall’.”  He   rose   from  his   chair   and  walked  towards  the  door.  Other  guys  quickly  jumped  off  their  seats  too,  and  followed  him.  Tanya  alone  remained.  She  stayed  in  her  chair  for  a  little  while,  then  finally  sighed  and  headed   towards   the  doors  without   a   single  word.  However,   unlike   the  others  that   had   exited   the   room   and  moved   quickly   to   the   left,   she   turned   to   the   right,  which  clearly  is  not  the  way  to   ‘Town  Hall’.  Somehow  I  feel  that  this  isn’t  over  yet.  Not  with  Tanya,  anyway.      However,  one  important  battle  was  won  today.  Four  of  the  development  managers  returned   back   where   they   belong   –   helping   their   teams.   I   took   a   few  minutes   to  decompress   while   disconnecting   my   laptop.   Suddenly,   Stephanie   appeared   in   the  doorway  –  all  shiny  and  happy,  like  you  are  at  your  12th  birthday  party.    “Ryan,  I  don’t  know  what  you  did  to  them,  but  they  all  ran  into  ‘Town  Hall’,  split  up  and  started  discussions  with  their  teams.  I’m  staggered,  Ryan.  Thank  you  so  much!”    “Yeah,  but  Tanya  didn’t  seem  to…”    “Oh,  don’t  worry  about  it  –  that’s  business  as  usual  for  her.  She’ll  stab  us  in  the  back  later  –  that’s  her  style.  Let’s  go  now.  Saheli  just  texted  me  that  something  is  going  on  there.  We  need  to  go  back.”    She  put  her  phone  in  her  pocket  and  added:  “You  know,  John  is  absolutely  rocking  it  there.   He’s   running   around   the   room   and   he’s   probably   met   with   every   team  already,   some   even   twice.   He’s   assigned   business   value   to   basically   all   teams’  objectives,  as  you  said,  just  a  relative  number  one  through  ten,  and  assisted  many  of  

Page 41: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   41  

them  in  determining  what   they  should  set  as  a  stretch  objective.   I  even  heard  him  talking  on  the  phone  earlier  today,  canceling  some  of  his  meetings,  saying  something  like   ‘I   need   to   be   here   today’,   ‘I   need   to   finish   this   first’   and   so   on…  And   then   he  simply  hung  up,”  Stephanie  smiled.      I  automatically  smiled  in  return,  but  realized  that  for  the  last  couple  of  days  I  hadn’t  been  smiling  much.  However,  all   this  time  Stephanie  had  been  around  and  despite  all   the   hardships   of   this   adventure,   she   was   the   one   human   being   I   was   so  desperately  glad  to  see  every  morning.  We  would  talk  through  the  plan  for  the  day,  discuss  problems  and  share  concerns.  Was  that  just  a  reflection  on  the  fact  that  she  will  probably  make  the  best  Release  Train  Engineer  I’ve  ever  known,  or  am  I  taking  the   Lean   tenet   of   Respect   for   People   to   a   whole   new,   somewhat   personal   level?  Concentrate,  Ryan,  this  adventure  is  not  over  yet.  It  will  be  over  when  the  business  owners  tell  us  they  are  happy  with  the  plan,  and  the  teams  confirm  they  can  deliver  the  value.      Stephanie   opened   the   door   and   the   scene   before   us   was   absolutely   stunning.   It  appeared   that   team   members   had   re-­‐shuffled.   Some   of   them   joined   two  development  managers   in   the   corner   by   the   window,   busily   crafting   something   I  hadn’t  seen  before.      “Hmm…”  Stephanie  looked  around  and  said:  “Looks  like  lot’s  of  the  team  members  are  working  now  in  their  peer  teams’  areas.  Dependencies,  I  would  guess.  And  what  is  that?”  she  pointed  at  the  corner  that  had  me  completely  puzzled.      Saheli   approached   us   and   pointed   to   the   same   corner:   “Guys,   I   don’t   know  what  happened,   but   about   thirty   minutes   ago   Brian   and   some   other   development  managers  literally  ran  into  the  room,  circled  a  time  or  two,  carefully  observing  the  team  plans,  and  then  started  offering  help  to  the  teams  that  needed  it.  They  quickly  noticed  that  the  Wolverines  had  some  dependencies  on  the  Spartans,  but  that  those  were  actually  based  on   false   timing  expectations.  Something   that  was  expected  by  the  Wolverines  in  Sprint  two  was  actually  to  be  delivered  only  in  Sprint  four  by  the  Spartans.   Then   Brian”   –   I   quickly   realized   that   he  was   the   one   leading   the  merry  company  back  to  Town  Hall  –  “Gathered  the  other  managers  and  the  Scrum  Masters  and  started  mapping  those  dependencies  for  the  Train.      “That’s   actually  what   they   are   doing   there   in   the   corner   as  we   speak   –   creating   a  dependency   board.   It’s   a   matrix   with   teams   as   rows   and   Sprints   in   the   PI,   as  columns.  There  are  stickies  of  two  colors:  blue  and  red.  Blue  stands  for  a  feature.  If  a  blue  sticky   is   in  a  particular  cell,   it  means   that   that   feature  will  be   finished  by   the  team  in  that  particular  Sprint.  But  other  teams  may  also  have  inputs  to  that  feature,  which  are  the  red  stickies  connected  to  the  blue  ones  with  the  string.  I  have  no  idea  where   Brian   found   the   string,   but   it   seems   like   they   know   what   they   are   doing.  Within  the  last  ten  minutes,  they  found  four  more  dependencies  between  teams  that  were   not   synchronized   in   time.   That’s  why   the   teams   are   now  working  with   one  another  –  they  are  re-­‐planning  for  those  dependencies.  

Page 42: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   42  

 

   Wow.   I   was   certainly   expecting   Brian   and   his   fellow  managers   to   come   and   help  their  teams,  but  I  was  not  expecting  miracles.  Meanwhile,  Brian  was  so  busy  there  at  the  board  that  he  hadn’t  even  noticed  us.      “But   that’s   not  why   I   texted   you,   Stephanie,”   Saheli  went   on.   “Team  Moria   has   no  stretch  objectives.  They  are  saying  they  don’t  have  any  choice  other  than  to  commit  to  100%  of  the  scope  that  they  have  on  their  sheets.”    “What   are   they   working   on?”   I   have   seen   such   things   a   couple   times   before,   but  every  time  it  was  for  a  different  reason.      “They  are  basically  working  on  legal  and  compliance  backlog  items.  Let’s  go  talk  to  them.”    “We  don’t  have  any  other  choice,”  said  the  Scrum  Master  of  Moria.  “If  we  don’t  finish  this  whole  scope,”  he  waived  his  hand  across  the  sheets  with  stickies,  “We  will  not  be  able  to  even  sell  most  of  our  perishable  goods,  nor  will  we  be  able  to  take  debit  and  credit  cards.  These  are  mandatory  things  to  do  and  we  just  have  to  roll  up  our  sleeves  and  make  it  happen.”          I  moved  a  little  closer  to  the  Sprint  sheets  to  see  the  numbers:  “But  those  Sprints  are  fully  loaded.  Some  are  even  over  capacity.”    “Yeah,  but  there  is  nothing  we  can  do…”  

Page 43: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   43  

 “Look.   The   fact   that   this   entire   scope   needs   to   be   delivered   doesn’t   automatically  make   it   possible.   The   urgency   of   this   scope   and   your   team’s   ability   to   deliver   are  practically  unrelated  things.  If  we  do  that,  if  we  simply  accept  this  plan  and  assume  that  the  team  will  do  miracles,  that  will  surely  lead  to  a  much  bigger  problem  later,  because  you  will  simply  fail  to  deliver.”    Stephanie  was  carefully  reading  the  team’s  objectives.  “I’m  sure  that  the  Argonauts  can  help  you  with  this  one,”  she  pointed  at   ‘Synchronized  tracking  of  used-­‐by  date  through  the  supply  chain’.  “Let’s  get  their  Product  Owner  over  here.  And  this  one  –  ‘In-­‐memory  encryption  mechanisms’  –  I’m  sure  it  can  be  addressed  too.”    In   the   next   twenty   minutes,   they   adjusted   the   plan   and   the   two   teams   that   had  offered  help  pulled  out  some  less  important  items  from  their  backlogs.  Team  Moria,  in   turn,   picked   some   future   research   items   as   their   stretch   objectives,  which   they  would  be  able  to  get  to  once  they  got  some  breathing  room.        The   second   breakout   was   approaching   the   finish   line.   Teams   were   still   making  adjustments,  but   it   seemed   like   they  were   finally  ready   to  present  decent  plans   to  the   business   owners.   The   last   short   break   before   the   final   plan   review   does   not,  however,  seem  to  distract  the  teams  from  their  planning  areas.  Most  kept  discussing  various  items,  pointing  at  the  Sprint  sheets.  Some  walking  to  the  dependency  board  and  back…            

Page 44: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   44  

VI.  SAVED  BY  DISBELIEF    “Accepted,”  said  John  merrily,  glancing  at  Saheli  and  Josh.  The  guys  nodded  and  gave  a  big  thumbs-­‐up  as  the  whole  room  applauded  the  product  owner  of  the  Spartans,  who  had  just  finished  presenting  their  final  plan.  The  presentation  went  pretty  fast,  generating   very   few   questions   from   the   business   owners.   Stephanie   asked   their  product   owner   to   gather   their   objectives-­‐   and   risk   sheets   and   bring   them   to   the  front  wall,  where  the  other  four  teams  had  already  attached  theirs  after  presenting  them   to   John.   Team  Erebor   presented   next,  which   took   their   product   owner   only  three  minutes.  Basically  every  team  was  trying  to  focus  their  presentation  on  the  PI  objectives,   and   the   business   value   assigned   to   those   objectives   by   the   business  owners.  And  at  the  end,  explicitly  going  through  the  stretch  goals  for  the  PI,  as  this  is  something  that  may  or  may  not  be  delivered  and  thus  needs  to  be  called  out.  Finally,  John  and  the  guys  approved  the  last  team  plan  and  the  front  wall  received  the  last  piece  of  the  program  plan.      

   “What  do  we  do  now?”  asked  John  while  rubbing  his  hands.      “Risks.  Program  risks,”  I  answered  John,  as  I  took  the  microphone  and  addressed  the  whole   group:   “Now,   there   is   one   more   thing   that   lies   between   us   and   the  commitment  –  risks.  We  will  go   through  all   the   teams’  risk  sheets  one-­‐by-­‐one  and  build  a  consolidated  risk  sheet  for  the  program.”    In  the  meantime  Stephanie  attached  four  new  flipcharts  to  the  windows  and  named  each  one  of  them:      

Page 45: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   45  

Resolved.    Owned.    Accepted.    Mitigated.      “Let’s   ROAM,”   she   said   and  picked   the   closest   risk   sticky   from   the  Argonauts   risk  sheet  and  read  it  out  loud  to  the  group:  “Class  dependency  mapping  for  VB.Net  to  C#  module   translation.”   She   raised   the   sticky  up   in   the   air   and  asked   for   someone   to  comment  on  it.  One  of  the  Argonauts  quickly   jumped  out  of  his  seat:  “We  have  the  tool  to  assist  us  in  the  initial  translation  of  the  VB.NET  module,  but  since  we  know  that  such  tools  are  not  perfect,  and  we  will  still  have  to  manually  check  a  lot  of  real-­‐time   dependencies   between   instances   of   classes,   we   would   like   to   have   some  automatic  way  to  validate  whether  the  dependencies  were  preserved  in  a  translated  C#   source   code.   Without   such   a   thing,   we   could   actually   be   slowed   down   quite  substantially…”    “I  know  some   tools,”  Todd,   the  architect,   raised  his  hand.   “I   can  show  you  at   least  two  of  the  tools  that  I  know,  but  you  guys  will  have  to  take  it  from  there.”    “Great,”   said   Stephanie.   She   put   Todd’s   name   on   the   sticky   and   moved   it   to   the  ‘Mitigated’  sheet  of  the  consolidated  program  risk  area.    She  picked  the  next  sticky  and  then  the  next…  More  and  more  stickies  appeared  on  the   program   sheets.   Some   weren’t   actually   that   risky,   as   clarified   by   other   team  members.  These  went  to  the  ‘Resolved’  sheet.  Others  we  could  do  nothing  about  and  those  were  put   in   ‘Accepted’.  She   finally  pulled   the   last  sticky   from  the   last   team’s  risk  sheet:    “Here’s  the  last  one  from  Ninjas.  Releasability.”  Stephanie  looked  on  the  other  side  of  the  sticky  but  did  not  find  any  further  detail.  “Very  specific,”  she  smiled  and  asked  for  the  author  of  that  sticky  to  elaborate…      “We  have  an  overly  formal  process  with  our  release  operations.  We  never  know  if  a  release   is  going   to  happen  smoothly  or   if  we  will  have  deployment   issues,   like  we  have  had  multiple  times  before.  If  we  keep  doing  it  this  way,  we  may  not  meet  the  dates,”  said  Ninjas’  Scrum  Master.      “I’ll  take  it,”  said  Brian.  “I  hear  what  you  guys  are  saying  and  I’m  not  too  crazy  about  the  brick  wall  between  you  and   the  deployment  operations   team.  We  are  going   to  change  this  and  I’m  taking  responsibility  over  this  item,”  said  Brian,  proudly  adding,  “Please  put  my  name  on  that  one.”      Looks  like  Brian  is  totally  into  his  role  as  a  leader…  an  enabler.      “Thank  you  Brian,  that  would  make  our  life  2.5  times  easier,”  said  the  Scrum  Master.      

Page 46: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   46  

Brian  was  absolutely  shining:  “My  pleasure…”      The  ROAM’ing   session  was  over.  We  decided   to   remove   the   ‘Resolved’   sheet   from  the  window  as  tracking  those  items  was  of  no  value.      

   Stephanie  handed  me  the  microphone  for  the  culmination  of  the  PI  planning  session  –  the  Team  Confidence  Vote.      “Thank  you  Stephanie.  You  guys  have  probably  noticed  that  the  key  to  this  planning  process   is   not   to   have   the   management   plan   for   you,   but   to   have   you   plan   for  yourselves,   to   have  you  manage  dependencies   and   risks.   It   is   not   enough   that   the  business   owners   approve   your   PI   objectives.   That   only   means   that   they   have  confirmed  to  you  your  understanding  of  what  they  want  you  to  build.  You  and  only  you,  can  commit   to   that  scope.  So  now  I’m  going  to  ask  every  team  to  do  a  simple  confidence   vote.   Each   team   member   will   have   to   do   fist-­‐of-­‐five   to   show   your  confidence   level   in   delivering   this   scope   in   ten   weeks.   If   the   average   is   three   or  more,   we   take   it   as   a   sufficient   confidence   level.   So,   Ninjas,   let’s   start   with   you.  Ready?  Three,  two,  one,  go!...    “Five,   four,   five,   five,   four,   three,   four.   Very   good.   I   think   that   deserves   some  applause.”      The  room  erupts  in  cheerful  applause.  The  next  team  vote  count  is  almost  all  fives,  as  is  the  next  one.      “Argonauts.  Ready?  Go…”  

Page 47: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   47  

 Wow.  We  have  a  one!  What?  A  one?      “We  need  to  hear  from  you  sir,”  Stephanie  makes  her  way  with  the  microphone  to  a  developer  who,  I  hope,  simply  misunderstood  the  process.  He’s  the  guy  I  had  noticed  walking  around  the  room  during  the  risk  management  session.  He  had  been  staring  at  other  teams’  plans  and  taking  notes.  Well,  let’s  see  what  he’s  got.      “I’m  a  developer  that  was  previously  on  the  maintenance  team  that  was  disbanded.  Now  I’m  with  Argonauts  and  I  brought  some  backlog  items  with  me  from  that  team.  You  can  see  those  in  our  plan  –  those  purple  stickies.”  He  pointed  to  the  far  corner  where  his  new   team’s  Sprint  plans  were  hung.   “Then   I   realized   that   some   routine  bug  fixing  that   I  used  to  do  myself  was  not  on  my  team  plan.   I  wondered  where   it  went.  I  started  walking  across  the  room  and  looking  at  other  teams’  purple  stickies.  What  I  found  was  that  not  only  is  that  bug  fixing  scope  nowhere  in  the  plan,  but  that  the  overall  amount  of  maintenance  work  across  the  teams  seems  to  be  considerably  below  our  prior  workload,  compared  to  when  I  was  part  of  the  maintenance  team.  It  was   actually   quite   easy   to   figure   out.   Two   days   ago   Ryan   taught   us   Normalized  Estimation,  which   is  basically  about   starting  with   the   same  basis   for  a   story  point  across  the  Train,  in  order  to  be  able  to  compare  different  backlog  items.  Something  doesn’t  add  up.  I  think  we  are  missing  a  lot  of  critical  maintenance  work,  guys.”        Brian   and   Todd   where   already   walking   around   the   room.   Other   team   members  stood  up  and  started   looking   in   their  plans.  Brian  was  making  notes  and  when  he  finished  his  first  pass  around  the  room,  said:      “He’s  right.  We  missed  a  whole  bunch  of  things  that  we  used  to  do.  We  have  to  re-­‐plan.”      In  the  next  thirty  minutes,  Brian  and  all  of  the  ex-­‐maintenance  team  members  made  a  quick  check  and  identified  the  missing  areas.  Meanwhile,  the  other  teams’  product  owners  worked  with  Saheli   and   Josh   to  determine  what   they   could  pull   out,   or   at  least  move  to  stretch,   in  order  to  make  room  for  more  maintenance  effort.  Finally,  when  the  adjustments  where  made,  all  the  teams  returned  to  their  tables.  The  walls  blossomed  with  a  comforting  purple  color.        “Almost   forty  percent  maintenance  missed?!”  Brian   took   the  microphone.   “Sergey,  thank  you  for  voicing  those  concerns,”  he  pointed  to  the  developer  who  had  spotted  the  problem.  “I  think  if  we  all  have  courage  and  attention  to  detail   like  Sergey  just  demonstrated,  Pacific  Express  has  a  great  chance  to  succeed.”    After  this  change  we  obviously  had  to  redo  the  confidence  vote,  but  this  time  all  the  teams  voted  quite  quickly  and  all  of  them  far  exceeded  the  required  minimum.      “Now,  guys,”   I   took  the  mic  again,   “we  have  dependencies  on  one  another,  and  we  have   to   frequently   synchronize   and   integrate   our  work.  We   have   to   operate   as   a  

Page 48: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   48  

Train,  not  just  as  individual  Agile  teams.  This  means  that  we  must  have  a  final  vote  as  a  Train  –  all  of  us.  If  you  have  any  concerns,  if  you  think  that  some  of  those  things  in  the  plans  we,  as  a  Train,  will  not  be  able  to  accomplish  –  now  is  the  time  to  say  it  out  loud.”    Everyone   raised   their   fist-­‐of-­‐five.   And   I   can   see   that   the   votes   are  well   above   the  three-­‐point  average.  Which  means  we  are  done.      “Congratulations,  Pacific  Express.  You  got  your  plan.  It  was  my  pleasure  to  assist  you  in  this  venture,  but  keep  in  mind  that  you  essentially  did  all  of  this  yourself.”    John  suddenly  picked  up  a  second  microphone  and  joined  me  on  the  stage:  “Ryan,  I  think  we   all   owe   you   one.   After   going   through   these   two   days   of   planning,   I   now  realize   how   much   we   accomplished   in   terms   of   alignment   and   gaining   a   true  understanding   of   what   we   are   building   here.   I   think   we   all   learned   a   lot   in   this  process.  And,   as   the  one   and  only   true   representative  of   the  business  here   in   this  room,   I   can   tell   you   that   I   am   really   happy  with   the   outcome.   I   think  we’ve   got   a  plan…”      

*  *  *    One  hour  later,  Stephanie  and  I,  tired  but  very  happy  with  the  result  of  the  session,  were  ordering  appetizers  at  Tapenade  restaurant  in  La  Jolla.  She  said  that  the  other  guys  were  not  able  to  join  us,  which  I  absolutely  believed  to  be  the  case.  At  least  this  would  let  us  have  an  honest  conversation  about  the  flow  of  value,  limiting  WIP  and  controlling  the  queue  length...  And  in  the  morning  I  will  take  off…            

Page 49: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   49  

EPILOGUE    It  was  a  gentle  afternoon  rain,  tapping  softly  on  my  porch  and  trying  to  seduce  the  entire  world  into  a  nap.  It  may  be  the  last  rain  of  the  season  –  whatever  falls  next  out  of   the  sky   is   likely  to   freeze  before   it   touches  the  ground.  The  grass  has  all   turned  yellow   and   my   garden   is   withered   –   almost   ready   for   the   harshness   of   winter.  Loaded  with   coffee,   I   was   busily  working   on   a   presentation   that  would   hopefully  expand  into  a  brand  new  training  course  some  day.  I  was  taking  a  longer  break  from  traveling:  being  on  the  road  all  the  time  leaves  no  time  to  invent  new  things  and  also  leads  to  the  desolation  of  my  garden.  Both  require  time  to  grow,  both  need  attention  and   effort   to   result   in   something   pleasing   to   the   eye.   There’s   not   much   growing  outside   this   time  of   the   year   –   only  dry   stems  poking  out   of   the   ground.  Only  my  little   spruce   is   full   of   resilience   and   life.   This   is   the   perfect   time   to   rearrange   the  flower-­‐beds  and  re-­‐pave  those  tiny  little  pathways  between  them.    I  was  finishing  the  current  slide  as  well  as  my  second  cup  of  coffee  when  the  phone  broke  the  silence  with  its  rhythmic  humming,  slowly  shifting  on  the  table  in  random  directions.  It  was  Stephanie  calling.    “Ryan,  how  are  you?   Is   this  good   time   to   talk?  Our   first  PI   comes   to   conclusion   in  about   two  weeks   and  we   need   you   to   help   us  with   the   program   Inspect  &  Adapt  workshop.  Sorry  for  such  short  notice,  we’ve  all  been  really,  really  busy…”      She  told  me  how  they  had  failed  to  integrate  the  system  early  on  and  how  that  led  to  a   failed  system  demo  after  Sprint  one.  Consequently,  Tanya  had  come  back  almost  immediately   with   a   lot   of   criticism   and   begun   busily   campaigning   among   the  executives  to  shut  down  SAFe  adoption.    “…But   here’s   what   happened,”   Stephanie   proceeded,   “She   faced   a   relentless  resistance  from  her  own  development  managers.  I  can’t  tell  you  how  many  conflicts  between  Tanya  and  Brian  I  personally  witnessed.  And  those  were  no  fun  to  listen  to,  trust   me.   But   Brian   and   other   development  managers   provided   assistance   to   the  teams  in  every  imaginable  aspect.  They  all  swarmed  around  the  integration  problem  and  guess  what?  In  Sprint  two,  we  witnessed  our  first  system  demo.  John  attended  and  was  as  happy  as  a  kid  in  a  candy  store,  asking  guys  to  run  this  and  to  click  that  and   ended   up   taking   over   the   keyboard.   Certainly   we   had   to   de-­‐scope   a   little   to  recover   from  our   sluggishness   in  Sprint  one,  but   it  was  a  good  demo  nonetheless.  Immediately   after   the   demo,   Brian   ran   around   the   entire   building   and   personally  thanked  every  Agile  team.  Since  then  we  have  had  a  fully  integrated  increment  every  Sprint.  Yesterday  was  our  last  system  demo  and  today  we  started  Sprint  five.      “We  need  you  very  badly  for  the  last  couple  days  in  this  Sprint,  where  we’ll  do  the  grand  demo  of  the  entire  PI  scope.  This  time  not  only  will  John  be  there,  but  other  guys   from   the   business   side   said   they  would   like   to   attend.   Also,   our   CIO  will   be  there.”  Stephanie  paused  for  a  couple  moments.  “You  know,  they  actually  redefined  

Page 50: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   50  

Tanya’s   position.   She   still   holds   that   job,   but   development   managers   no   longer  report  to  her.  Also,  Brian  was  promoted  to  a  director  of  development,  which  is  a  de  facto  replacement   for  Tanya.  Anyway,  we  see  her   less  and   less  now  which   is  a  big  win  for  everyone,”  she  chuckled.  “So…  Can  you  come?”    “Yes.  I  will.”    “Awesome   sauce!”   She   said,   bursting   into   laughter.   “I   have  one  other  question   for  you  though.”    “Yes…”    “We  have  only  now  begun  to  realize  that  we  achieved  a  qualitative  shift  in  progress  tracking  –  by  producing  fully  integrated  increments  of  the  solution  every  two  weeks.  But  we  also  realized  that  now,  as  we  are  ending  the  PI,  we  don’t  have  a  single  metric  to  show  how  successful  it  was  for  the  Train  as  a  whole,  or  to  see  how  good  we  did  as  individual  teams.  We  will  certainly  have  the  grand  PI  demo,  but  that  doesn’t  allow  us  a  finer  perspective  on  our  accomplishments.  And  it  is  a  pity  that  we  have  only  now  realized  that  we  don’t  have  a  single  metric  at  our  disposal,  here  at  the  end…”    “Oh,  that  is  so  not  true.  You  guys  have  a  metric  already.”    “No  we  don’t.”    “Sure  you  do.  What’s  on  all  those  PI  objective  sheets  the  teams  have?”    “Well,   committed  objectives,   stretch  objectives  and  business  value   for   each  one  of  them.”    “There   you   go!   But   that   is   planned   business   value.   That’s   what   business   owners  ideally  expect  from  the  teams.  Now  is  the  time  to  let  John  and  his  guys  fill  out  actual  business   value,   as   they   perceive   it   based   on   the   upcoming   PI   demo.   Imagine   this.  You  will  have  two  columns  on  every  sheet:  planned  and  actual.  Each  team  will  then  add  them  up  and  get  two  numbers:  total  planned  and  total  achieved  business  value.  Your  metric  is  the  ratio  of  the  two.  The  numbers  then  can  be  consolidated  and,  voila,  you’ve  got  your  metric  for  the  Train.  By  the  way,  if  you  want  to  measure  anything  at  all  for  the  Train,  you  want  to  measure  this.  And  this  metric  has  some  magic:  it  gives  you  an  idea  of  productivity,  but  expressed  in  the  most  relevant  currency  possible  –  percentage  of  achieved  business  value.  And  who  defines   those  numbers?  Those   in  the  organization  who  ultimately  determine  value.  You  can  never  be  wrong  when  you  directly  ask  the  business  instead  of  speculating  on  your  own.”    “Ryan  this  is  awesome.  I’m  so  looking  forward  to  seeing  you…”    Seems  like  the  Train  is  acquiring  momentum.  I’m  actually  glad  they  failed  their  first  system   demo   in   Sprint   one.   At   least   now   they   clearly   understand   that   our   initial  

Page 51: Pacific Express - rev 3 · © %2014%Scaled%Agile,%Inc.% 5 This%is%getting%more%and%more%interesting.%“Josh,%if%they%are%Sprinting,%don’t%they%also% do%Sprint%demo%in%the%end?”%Iasked.%

©  2014  Scaled  Agile,  Inc.   51  

sessions,   including   the   PI   planning,  were   not   the   transformation   itself.   There   is   a  reason   why   it   is   called   Quickstart.   In   fact,   it   is   only   the   beginning   of   a   learning  journey   that  knows  no  end.  But   there   is  something  different  about  Pacific  Express  now.  They  have  learned,  although  it  was  through  much  pain,  how  to  deliver  end-­‐to-­‐end   value.   That   should   open   their   eyes   to   all   other   things.   Most   programs   can’t  improve  because  they  are  blind  to  the  things  that  really  matter  –  they  don’t  see  the  system   as   a   whole.   Neither   the   one   they   are   developing,   nor   the   one   they   are  themselves  part  of.  This  Train,  it  seems,  succeeded  with  this  first,  smallest  but  most  important  step  –  they  acquired  the  bigger  picture  view.      The  rain  has  almost  stopped.  A  few  drops  can  still  be  seen  here  and  there,  flaring  in  the  air,  as  the  clouds  start   to  give  way  to  the  magnificent  azure  sky.   I  can’t  sit  and  stare  out  in  my  window  any  longer  –  I  need  to  go  outside  and  feel  the  last  touch  of  the   rain,   the   kindness   of   the   sun   and   the   playful   breeze   from   the  mountains   that  wear  their  stunning  white  mitres  quite  prematurely  this  year.    


Recommended