Enterprise Systems and Innovations

Benjamin Engelstätter

ZEW Mannheim

CoInvest Lisbon, Portugal

March, 18 - 19 2010

Brief IntroductionEnterprise Systems (ES)

Software to control, manage and support business processes

Three Main Branches• Customer Relationship Management (CRM): Front Office• Enterprise Resource Management (ERP): Middle Office• Supply Chain Management (SCM): Back Office

Additional Types• Technical Software (CAx)• MES, PLM, …

Market• 39 billion USD for complex enterprise systems in 2008, 1.9 Bil. Euro in Germany• Market for large firms is satisfied, SMEs are now focused• especially ERP, SCM and CRM spread out worldwide

Enterprise Systems and Innovations

Enterprise Resource Planning • standardizes complex interfaces and automates financial transactions • collects and updates firm intern data in real-time

Supply Chain Management• coordinates flow of information, materials and finances along the value chain• improves operational and business planning with real-time planning capabilities

Customer Relationship Management• provides a firm-wide centralized database of customer information• offers a complete view of customer needs and wants

Possible Effects on Innovations• SCM & ERP identify bottlenecks and shortages• generated databases provide exact information facilitating process enhancements• CRM database can be used as information source for product innovations

Contribution

Effects on Innovation

• First empirical evidence of the impact of adopting any of the

three main enterprise systems on firms’ innovational

performance

• process as well as product innovations are concerned

• 1st Step: Revealing impacts of enterprise systems on

probabilities to innovate

• 2nd Step: Revealing impacts of ES on number of realized

innovations

LiteratureDirect Effects on Innovation

• ERP facilitate the building of business innovations (Shang & Seddon, 2000)

• customer preferences retrieved via CRM improve innovational success

(Joshi & Sharma, 2004)

• ES allow people to be more innovative (Davenport, 1998)

Indirect Effects on Innovation

• business units more innovative if in central network position (Tsai, 2001)

• more innovation through upstream /downstream contacts to suppliers and

customers (Chriscuolo et al., 2004)

• organizational flexibility leads to increased innovative activity (Hempell & Zwick,

2005)

→ ES offer advantages in all categories and might foster innovational performance

Research Methodology

Innovation/Knowledge Production Function

output of innovation process represents result of several research linked inputs

(1) zi* = Xi’β1 + IDi’β2 + ESi’β3 +εi zi = 1 if zi* ≥ 0; zi = 0 otherwise

Number of innovations

(2) yi* = Zi’λ1 + IDi’λ2 + ESi’λ3 +γi yi= yi* if zi = 1; yi= 0 if zi = 0

Variableszi – Product/Process Innovation

yi – Number of Product/Process

Innovation

Zi / Xi – determinants affecting innovation

IDi – control dummies for industry sector

ESi – Enterprise Systems in use

εi / γi – standard error term

Estimation Procedure

Procedure

• Maximum Likelihood

• count data corner solution with 2-part model

• 2 alternatives:

- Hurdle model

- Zero-inflated model

• both allow for separate treatment of zeros and strictly positive

outcomes

• 2 possible distributions:

- Poisson

- Negative-binomial

Possible Models

Hurdle model

• reflecting 2 stage decision making process

• each part model of one decision

• f1(·) determining zeros, f2(·) determining positive counts

• both parts functionally independent

• 1st part uses full sample, 2nd only positive count observations

Zero-inflated Model

• f1(·) determining zeros, f2(·) determining positive counts

• 2 types of zeros:

- one type arising from binary process

- other type is realization of count process (when binary process takes on

1)

1

12

2

(0)

1- (0)( )

1- (0)

f

ff y

f

if 0,

if 1

y

y

f (y) =

1 1 2

1 2

(0) + 1- (0) (0)

1- (0) ( )

f f f

f f y

if 0,

if 1

y

y

f (y) =

Database

ZEW ICT Survey

• computer-aided telephone survey

• specific focus on diffusion and use of ICT in German companies

• one recent ICT topic specifically covered each wave

• each wave contains about 4000 firms with 5+ employees

• seven branches of manufacturing, seven selected service sectors

• five waves (2000, 2002, 2004, 2007, 2010)

• waves of 2004 and 2007 used in current analysis

VariablesTable 1: Summary statistics

Variable MeanStd. Dev. DV2

Software Use 04 Mean DV2

process innovations 04-06 0.635 yes no software 0.231 yes

number of process innovations 3.103 4.057 ERP 0.635 yes

product innovations 04-06 0.600 yes SCM 0.434 yes

number of process innovations 5.005 9.639 CRM 0.524 yes

process innovations last period 0.755 yes all three 0.275 yes

product innovations last period 0.654 yes

labor1 213.0 636.4 Additional control variables

share of computer workers 0.469 0.329 working hours3 0.704 yes

share of highly skilled workers 0.226 0.259 job rotation 0.191 yes

share of medium skilled workers 0.557 0.262 quality circles 0.425 yes

ISO certificated 0.444 yes own cost units4 0.386 yes

East Germany 0.267 yes workgroups5 0.623 yes

Number of Observations 989

Notes: 1 Labor is measured in total number of employees. 2 Dummy variable. 3 Accounts for working hours. 4 Units with own cost and result responsibilty. 5 Self dependent workgroups; Source: ZEW ICT survey 2004, 2007. Own calculations.

Descriptive Evidence and Model Selection

Table 2: Descriptive analysis

No system

All systems ERP SCM CRM

recent process innovator 0.478 0.790 0.726 0.767 0.726

number of process innovations mean

1.783(2.774)

4.039(4.252)

3.764(4.402

)

3.935(4.424

)

3.590(4.274

)

recent product innovator 0.408 0.728 0.667 0.700 0.681

number of product innovations mean

2.142(5.045)

7.167(10.82)

6.002(10.15

)

6.859(10.87

)

6.197(10.45

)

Notes: Standard Errors in parentheses. Source: ZEW ICT survey 2004, 2007 and own calculations.

Process innovations Product innovations

Vuong-Test 6.050*** 7.830***

Llhd.-ratio Test

449.220*** 2864.020***

Notes: *** p<0.01, ** p<0.05, * p<0.1; Source: ZEW ICT survey 2004, 2007 and own calculations.

Table 3: Model selection

Zero-inflated neg. bin. model selected in both cases

Results – Process Innovations

Specification (1) Specification (2)

ProbitModel

Neg. Bin.Model

ProbitModel

Neg. Bin.Model

ISO certification -0.270**(0.140)

0.230**(0.092)

-0.137(0.149)

0.196**(0.095)

Process innovations last period -0.377**

(0.130)0.262***(0.098)

-0.310**(0.139)

0.242**(0.105)

Enterprise Resource Planning -0.141

(0.136)0.282***(0.105)

-0.110(0.138)

0.279***(0.105)

Supply Change Management -0.318**

(0.137)0.064

(0.089)-0.284*(0.153)

0.059(0.096)

Customer Relationship Management

-0.181(0.133)

-0.068(0.094)

-0.198(0.141)

-0.103(0.095)

ControlsIndustry, East,Size, Workforce

Char

Industry, East,Size,

Workforce Char

Industry, East, Size, Workforce

Char, Org Factors

Industry, East, Size, Workforce

Char, Org Factors

Number of Observations 890 (547 non-zero, 343 zero)

Table 4: Determinants of the number of process innovations, zero-inflated neg. bin. estimates

Notes: *** p<0.01, ** p<0.05, * p<0.1; robust standard errors in parentheses. Overdispersion coefficient alpha highly significant (not reported). Source: ZEW ICT survey 2004, 2007 and own calculations.

Marginal Effects and Robustness Checks - Process Innovations

Spec. 1 Spec. 2

overall marg. effect ERP 0.972***

(0.291)0.923***(0.294)

overall marg. effect SCM 0.624**

(0.292)0.556*(0.300)

overall marg. effect CRM 0.057

(0.288)-0.024(0.288)

Table 5: Marginal Effects (short-term)

Spec. 1 Spec. 2

overall marg. effect ERP 0.856**

(0.360)0.913**(0.376)

overall marg. effect SCM 0.415

(0.457)0.271

(0.447)

overall marg. effect CRM 0.122

(0.391)0.019

(0.383)

Table 6: Marginal Effects (medium-term, ES use 02)

Notes: *** p<0.01, ** p<0.05, * p<0.1; robust standard errors in parentheses. Source: ZEW ICT survey 2004, 2007 and own calculations.

Controls

0.013(0.015)

CRM

0.031**(0.015)

SCM

0.011(0.014)

ERP

0.030**(0.013)

Product innovations last period

0.120***(0.040)

Share of high skilled workers

workforce characteristics, ISO, former process innovator n. s.

Dependent variable: R&D spending in share of total sales in 2006 (OLS)

Table 7: R&D spending and ES usage

Industry, East,Size, Org Factors

Results – Product Innovations

Specification (1) Specification (2)

ProbitModel

Neg. Bin.Model

ProbitModel

Neg. Bin.Model

ISO certification -0.364**(0.172)

-0.135(0.142)

-0.345**(0.174)

-0.109(0.138)

Product innovations last period -1.067***

(0.152)-0.010(0.182)

-1.051***(0.151)

-0.031(0.177)

Enterprise Resource Planning -0.085

(0.170)0.027

(0.150)-0.144(0.173)

-0.091(0.155)

Supply Change Management 0.135

(0.182)0.116

(0.140)0.153

(0.180)0.070

(0.134)

Customer Relationship Management

-0.296*(0.161)

0.084(0.129)

-0.326**(0.163)

0.012(0.126)

ControlsIndustry, East,Size, Workforce

Char

Industry, East,Size,

Workforce Char

Industry, East, Size, Workforce

Char, Org Factors

Industry, East, Size, Workforce

Char, Org Factors

Number of Observations 886 (490 non-zero, 396 zero)

Table 8: Determinants of the number of product innovations, zero-inflated neg. bin. estimates

Notes: *** p<0.01, ** p<0.05, * p<0.1; robust standard errors in parentheses. Overdispersion coefficient alpha highly significant (not reported). Source: ZEW ICT survey 2004, 2007 and own calculations.

Marginal Effects - Product Innovations

Specification 1

Specification 2

overall marg. effect ERP 0.345

(0.728)-0.084(0.757)

overall marg. effect SCM 0.238

(0.730)-0.035(0.704)

overall marg. effect CRM 1.156*

(0.675)0.866

(0.654)

Table 9: Marginal Effects (short-term)

Specification 1

Specification 2

overall marg. effect ERP -0.385

(1.089)-0.611(1.139)

overall marg. effect SCM -0.390

(0.984)-0.344(0.973)

overall marg. effect CRM 1.544

(1.130)1.522

(1.137)

Table 10: Marginal Effects (medium-term, software use 02)

Notes: *** p<0.01, ** p<0.05, * p<0.1; robust standard errors in parentheses. Source: ZEW ICT survey 2004, 2007 and own calculations.

ConclusionMain Results

• ERP+SCM positively impacts number of process innovations

• SCM usage lowers probability of being a non-innovator in case of process

innovations

• both results stable for short and medium-run

• CRM users face a higher probability to product innovate (only short-term based)

Implications

• manager should not only focus on possibly huge costs and expected fast

evolving performance benefits when purchasing or upgrading ES

• increased process innovational performance via SCM and ERP might even

reduce costs

• product innovations realized based on CRM data might increase financial

performance via opening up new markets