Questions:
1. Service Innovation: Did your rm introduce a new or essentially improved service during the last 12 months?
2. Enterprise Software: Which types of enterprise software are used within your rm?
The ZEW quarterly business survey among service providers of the information soci-ety includes the following industries (NACE Rev. 1.1 Codes of European Commu-nity in parenthesis): software and IT services (71.33.0, 72.10.0-72.60.2), ICI-specialised trade (51.43.1, 51.43.3-3.4, 51.84.0, 52.45.2, 52.49.5-9.6,), telecommunication services (64.30.1-0.4), tax consultancy and accounting (74.12.1-2.5), management consultancy (74.11.1-1.5, 74.13.1-3.2, 74.14.1-4.2), architecture (74.20.1-0.5) technical consultancy and planning (74.20.5-0.9), research and development (73.10.1-73.20.2) and advertising (74.40.1-0.2). Table 3 shows the distribution across industries in the sample of 336 observations.
Table 3: Distribution of Industries in the Sample
Industry Observations Percentage
software and IT services 43 12.80
ICT-specialised trade 33 9.82
telecommunication services 13 3.87
tax consultancy and accounting 56 16.67
management consultancy 37 11.01
architecture 54 16.07
technical consultancy and planning 34 10.12
research and development 38 11.31
advertising 28 8.33
sum 336 100
Source: ZEW Quarterly Business Survey among service providers of the information society, own calculations.
Table 4: Probit Estimation Results: Coecient Estimates
dependent variable: dummy for service innovation
(1) (2) (3) (5)
sector-specic software −0.151 0.078 −0.036 −0.022
(0.164) (0.191) (0.226) (0.277)
customised software 0.633∗∗∗ 0.630∗∗∗ 0.767∗∗∗ 0.586∗∗
(0.144) (0.166) (0.194) (0.234)
log. rm size 0.139∗∗ 0.082 0.071
(0.060) (0.072) (0.087)
employees 30 − 55 years −1.445∗∗∗ −0.629
(0.554) (0.667)
constant term −0.426∗∗∗ −0.410 1.794∗ 0.520
(0.156) (0.667) (0.929) (1.119)
observations 336 298 240 179
pseudo R2 0.046 0.103 0.147 0.206
Signicance levels: ∗: 10%,∗∗: 5%,∗∗∗: 1%.
Reference categories: competitors 6-20, unqualied employees, employees <30 years.
Table 5: Probit Estimation Results: Average Marginal Eects, Reduced Sample
dependent variable: dummy for service innovation
(1) (2) (3) (4)
sector-specic software −0.040 0.007 −0.001 −0.003
(0.079) (0.083) (0.084) (0.081)
customised software 0.290∗∗∗ 0.250∗∗∗ 0.239∗∗∗ 0.182∗∗
(0.071) (0.074) (0.074) (0.075)
employees 30 − 55 years −0.285 −0.184
(0.199) (0.194)
pseudo R2 0.071 0.145 0.157 0.206
Signicance levels: ∗: 10%,∗∗: 5%,∗∗∗: 1%.
Reference categories: competitors 6-20, unqualied employees, employees <30 years.
Table 6: Summary Statistics of Reduced Sample
Variable Mean Min. Max. N
service innovation 0.357 0 1 179
sector-specic software 0.754 0 1 179
customised software 0.424 0 1 179
number of employees 39 1 449 179
log (number of employees) 2.672 0 6.107 179
rm age 20 4 108 179
log (rm age) 2.849 1.386 4.682 179
0-5 competitors 0.229 0 1 179
6-20 competitors 0.329 0 1 179
more than 20 competitors 0.441 0 1 179
share of employees working with PC 0.799 0.01 1 179
share of highly qualied employees 0.424 0 1 179
share of medium qualied employees 0.153 0 1 179
share of low qualied employees 0.373 0 1 179
share of employees younger than 30 years 0.181 0 1 179 share of employees between 30 and 55 years 0.652 0 1 179
share of employees older than 55 years 0.159 0 1 179
prior service innovation 0.418 0 1 179
prior process innovation 0.418 0 1 179
East Germany 0.376 0 1 179
Source: ZEW Quarterly Business Survey among service providers of the information society, own calculations.
Table 7: Distribution of Industries in the Sample
Industry Observations Percentage
Source: ZEW Quarterly Business Survey among service providers of the information society, own
3 Why Adopt Social Enterprise Software? Impacts and Benets
∗Abstract
This paper explores the performance impacts and benets of the adoption of social enterprise software (SES). SES forms a nested innovation, given that its adoption requires an already established infrastructure of information and communication technology. To control for induced sample selection, we use a two-step estimation procedure. Based on German rm-level data our results conrm that rms which use business-to-business (B2B) e-commerce applications are more likely to adopt SES. The estimated correlations also provide weak evidence for complementarity between B2B e-commerce and SES. We show that two measures of rm performance, i.e. sales and labour productivity, are highest for rms using SES and B2B e-commerce applications in conjunction.
Keywords: adoption of ICT, social enterprise software, nested innovation, complemen-tarity
JEL-Classication: D00, L10, O31
3.1 Introduction
In recent years, social software, e.g. wikis, blogs, microblogs or social networks, has increasingly appeared in both public dialogues and press releases. Social software is already extensively used in private households and increasingly adopted by rms. As for rms, more than 80 percent of the top 100 companies in the Fortune 500 maintain a presence on social network sites (Gartner 2012). However, currently a new type of business software is emerging, which interrelates recent social software and rms' es-tablished enterprise systems, e.g. Enterprise Resource Planning (ERP), Supply Chain Management (SCM), Customer Relationship Management (CRM). This so called social enterprise software (SES) links rms' enterprise software systems and social software applications as in, e.g., social CRM solutions.
Overall, SES oers benets in the areas of customer (B2C) and business-to-business (B2B). For B2C, SES supports tracking data from customer surveys, customer feedback, reviews or user proles on social networks or blogs, thereby enabling rms to identify new customers, new market segments and observe recent trends. SES packages additionally feature various communication channels allowing for a two-way interaction between companies and their customers, by oering them a direct channel for providing their feedback. With specic customer data collected and direct customer interaction SES might facilitate the development of new products as it allows rms to observe customer tastes and build up meaningful customer proles.
In the B2B segment benets of SES emerge in form of enhanced collaboration and communication as employees and partners are connected in a way they can exchange information fast and freely. Stakeholders have real-time access to all areas they are in-terested in and can monitor and directly access interactions and inputs of others. SES further enhances process management and knowledge sourcing as knowledge consumers, like sales teams, can access information from knowledge providers, like product develop-ers, in real-time saving time for each employee. With partners and clients of the utilising
rm connected, the benets of SES are not limited to the boundary of the rm but may spread on to other business contacts.
Up to now there are no empirical studies on the emerging phenomenon of SES although these software packages began to come up in 2008 (Chess Media Group 2010).
Con-evidence about determinants of its adoption and its potential impacts on rm perfor-mance is still lacking. Also, up to now it is still unclear which distribution segment, i.e.
B2B or B2C, benets the most once SES is adopted.
We aim at lling this research gap by empirically evaluating the determinants of SES adoption and exploring its impacts on rm performance. In the analysis, we distinguish between benets and impacts in the area of B2B and B2C. Our analysis is based on a unique database consisting of German manufacturing and service rms. Since SES re-quires a rm to rst adopt particular information and communication technology (ICT) before it can upgrade them to SES, it represents a so-called "nested innovation" (Green-stein and Prince 2007). This "nested" structure induces sample selection which has to be taken appropriately into account in the estimation procedure.
Our study adds to the empirical literature in a number of ways. To our knowledge we are the rst to explore the performance impacts of most recent SES and investigate the determinants of its adoption. Second, considering that ICT might act as complements (Aral et al. 2012) or even substitutes (Kretschmer et al. 2012) in their impact on perfor-mance our results oer a weak test for complementarity based on correlations between the usage of established ICTs, i.e. B2B e-commerce applications, and the adoption of most recent ICTs, i.e. SES. Third, our paper presents a valid empirical method with which to model the data generating process in the case of a "nested innovation", i.e. the probit with sample selection (Berinsky 2004; Gourieroux and Jasiak 2007)
Our results show that rms using B2B e-commerce applications are more likely to adopt SES. B2C e-commerce applications fail to impact the adoption decision. The estimated correlations also provide weak evidence for complementarity between B2B e-commerce applications and SES. Concerning impacts on rm performance, we show that mean sales and labour productivity are highest for rms using SES and B2B e-commerce applications in conjunction.
The paper proceeds as follows: Section 3.2 summarises the empirical literature of ICT, its complementarities and performance impacts and explains SES, its classication as a "nested" ICT innovation and its benets. Section 3.3 presents the dataset whereas Section 3.4 highlights the empirical model. Section 3.5 provides a detailed explanation of
results and additional robustness checks used to clarify the validity of the results are presented in Section 3.6. Finally, Section 3.7 concludes.