Post on 21-May-2015
transcript
MULTI-CHANNEL ATTRIBUTION TRENDS & BEST PRACTICES FOR B2B MARKETERSChris Bailey
chris@chrisbaileyworks.com
Agenda
1. RESEARCH APPROACH2. WHY ATTRIBUTION MATTERS3. INDUSTRY TRENDS4. FUTURE TRENDS
RESEARCH APPROACH
Sources & Methodology
Key Sources
MethodologyResearched and reviewed analyst reports and
blogpostsCompiled and analyzed dataDuring time span of March – April 2013
WHY ATTRIBUTION MATTERS
“ Because marketers struggle with understanding the true performance of channels, they lack visibility into where to make the proper marketing investments…
By employing an attribution approach, organizations can better understand how their offline and online channels are really affecting the bottom line.
Forrester Research“
INDUSTRY TRENDS
8
State of Attribution - 2012
N = 86 (Note: multiple responses accepted)Source: Forrester Research: Q2 2012 Interactive Marketing Executive Panel Survey
Advanced Statistical Attribution
Multi-Touch Attribution
Single Attribution
No Tracking
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%
9%
34%
26%
47%
Multi-Touch Attribution42%
Single Attribution27%
No Tracking32%
State of Attribution - 2013
n = 243Source: BtoB Magazine: The Evolving B2B Purchase Process, 2013
What Does This Mean?
Nearly 60% of B2B Marketers still use either single attribution or do not
track attribution at all.
ATTRIBUTION MODELS
Attribution Models
Single AttributionFirst touchLast Touch
Multi-Touch AttributionFirst+Last TouchEven DistributionTime Decay
Advanced Statistical Attribution
Single Attribution: First Touch
Touch 1Search Ad
100%
Touch 2Asset
Download0%
Touch 3Email0%
Touch 4Webinar
0%
Conversion Event
How it worksAllocate all revenue credit to first touch.
Pros• Easy to implement.• Good insight into early stages of revenue
cycle.• Insight into “investment per” lead metrics.
Cons• Doesn’t account for influence of
subsequent touches.• Too much credit to lead gen programs.
Not enough to nurturing touches or Sales contributions.
Single Attribution: Last Touch
Touch 1Search Ad
0%
Touch 2Asset
Download0%
Touch 3Email0%
Touch 4Webinar100%
Conversion Event
How it works:Allocate all revenue credit to the last touch.
Pros• Easy to implement.• Insight into “investment per” lead metrics.
Cons• Doesn’t account for influence of prior
touches.• May encourage cutting of activities and
resources which do impact conversion.
Multi-Touch Attribution: First+Last Touch
Touch 1Search Ad
50%
Touch 2Asset
Download0%
Touch 3Email0%
Touch 4Webinar
50%
Conversion Event
How it works:Sometimes called Position-based Modeling. Allocate revenue credit evenly between first and last touch.
Pros• Fairly easy to implement.• Helpful for longer revenue cycles.• Insight into first campaign (drive
awareness) and last program (drive conversion).
Cons• Doesn’t account for influence of middle
touches.• Can be easily skewed by large deals or
long sales cycles.
Multi-Touch Attribution: Even Distribution
Touch 1Search Ad
25%
Touch 2Asset
Download25%
Touch 3Email25%
Touch 4Webinar
25%
Conversion Event
How it works:Sometimes called Linear Modeling. Automatically allocate an even percentage across all touch points leading to conversion.
Pros• Incorporates nurturing touches as well as
lead generation.• Helpful for longer revenue cycles with
many touches.• Stops in-fighting over who gets credit.
Cons• Risk of over-crediting low impact touch
points.• Dilutes possible importance of high
impact touch points.
Multi-Touch Attribution: Time Decay
Touch 1Search Ad
5%
Touch 2Asset
Download10%
Touch 3Email25%
Touch 4Webinar
60%
Conversion Event
How it works:Touch points closest in time to the conversion receive more credit.
Pros• Incorporates nurturing touches as well as
lead generation.• Helpful for longer revenue cycles with
many touches.• Focuses on all contacts associated with
deal, not just first.
Cons• May artificially inflate the importance of
latter touches.
Advanced Statistical Attribution
Touch 1Search
Touch 2Asset
Download
Touch 3Email
Touch 4Webinar
10%Conversion
How it works:Find all examples of one sequence – both converting and non-converting – and calculate conversion rate. Next, find similar sequence that excludes one touch from prior sequence and calculate conversion rate. Compare conversion rates and determine weight of missing touch.
Pros• Less guesswork helps drive better
decisions on where to devote resources.• Offers greater insight into importance of
specific touches within conversion sequence.
Cons• Likely requires specialized technical
tools and statistical expertise.• Requires large amounts of data to
generate statistical validity.
Touch 1Search Ad
Touch 3Email
Touch 4Webinar
6%Conversion
Attribution challenges
While technology has helped, challenges still remain in how we account for:
• Earned media • Brand awareness• Offline connections
FIVE STEPS TO ATTRIBUTION SUCCESS
What is the purpose of attribution to your revenue goals?What do you want to learn through attribution modeling?
#1. Clarify objectives
Which one will best address your objectives?
More complex models often come with increased costs…which model fits your budget?
#2. Don’t choose randomly
What data will you need to feed your model?If you don’t already have this data, how will you collect it?
#3. Know your data
How will you measure outcomes and analyze data?How will you learn from the data and make changes?
#4. Start experimenting
What’s your plan for moving forward on the maturity continuum?
#5. Seek constant improvement
THANK YOUChris Bailey :: 2013
chris@chrisbaileyworks.com
Sources
Forrester Research: The Forrester Wave: Cross-Channel Attribution Providershttp://blogs.forrester.com/tina_moffett/12-04-30-just_published_the_forrester_wave_cross_channel_attribution_providers
MarketingProfs: Five Marketing Mistakes You Can Avoid by Using Cross-Channel Attributionhttp://www.marketingprofs.com/articles/2010/3440/five-marketing-mistakes-you-can-avoid-by-using-cross-channel-attribution
Marketo: How to Measure ROI of Your Marketing Programshttp://blog.marketo.com/blog/2013/03/how-to-measure-the-roi-of-your-marketing-programs.html
Avinash Kaushik: Multi-Channel Attribution: Definitions, Models and a Reality Checkhttp://www.kaushik.net/avinash/multi-channel-attribution-definitions-models/