Ambidexterity and Open Innovation in Small and Medium Sized Firms (SMEs) Andy Cosh & Joanne Jin Zhang Centre for Business Research, University of Cambridge Presented at Open innovation: new insights and evidence conference Imperial College London, 25 th -26 th June 2012
Introduction Firms have increasingly relied on external knowledge to promote innovation and performance; and they proactively collaborate with external partners throughout the innovation process, engaging in Inbound activities -- purposeful inflow of external knowledge. Outbound activities -- purposeful outflow or externalizing internally owned knowledge and technology (Enkel, Gassmann & Chesbrough, 2009). Source: H. Chesbrough
Research Objectives Yet, literature tends to look at either inbound or outbound in isolation. So how different forms of openness can be combined? Performance implications? (Dahlander & Gann, 2010) Scarce research on SMEs (a few exceptions e.g. van de Vrande et al 2009; Lee et al, 2010; Vanhaverbeke, 2012). SMEs face different challenges from larger firms in the innovation process. SMEs are key in the innovation system, and only mutual benefits can sustain the OI model. While there are some anecdotal evidence suggesting combining Inbound and outbound activities may lead to superior performance among larger firms, it is not clearly both theoretically and empirically if it works for SMEs? (Lubatkin et al., 2006) RQ: Do SMEs benefit from various forms of openness - engaging in inbound, outbound and both activities? if so, how?
External ambidexterity in open innovation Following March (1991) and others (Koza & lewin, 1998; Rothaermel & Deeds, 2004), we regard Inbound (seeking new knowledge) as exploration: the pursuit of new knowledge, of things that might come to be known Outbound (leveraging existing knowledge) as exploitation: the use and development of things already known. March (1991) argues that firms need to engage in both exploration and exploitation activities to achieve long term success. While earlier literature portrays the trade-off between Er and Ei as insurmountable, recent studies suggest that ambidexterity may help firms solve this dilemma. Internal vs. external ambidexterity We conceptualize ambidexterity in the OI process as simultaneous pursuit of inbound & outbound activities --emphasises a dynamic model of external knowledge process Shift from achieving ambidexterity internally to externally. SMEs face different challenges from larger firms in the OI process Rarely have the capacity to cover the whole process, so tend to be more open (OI is not new to SMEs) Lack of resources and capabilities to identify and absorb external knowledge Constraints in value capturing (weak in IP protection) So ambi might do SMEs harm as some have suggested (Ebben & Johnson, 2005).
Inbound, outbound & performance in SMEs Inbound /exploration - performance Motivated to discover something new; Returns are distant in time and more vulnerable; Promote innovativeness (e.g. Laursen & Salter, 2006). Outbound /exploitation - performance Promotes short term commercial performance Motivated to access complementary assets and generate rents through economies of specialization (Teece, 1986) Returns are more predictable, and promotes efficiency Generate additional revenue Market knowledge gained from partners in collaboration leads to new market opportunities. May also promote innovativeness such as selective free revealing and dual licensing (e.g. Wikipedia, Linux etc). H1a: Engaging in inbound activities is positively related to firm s innovation performance. H1b: Engaging in outbound activities is positively related to firm s innovation performance and sales growth.
Two dimensions of ambidexterity To achieve long term success, firms need to engage in both Er and Ei; but there is question of whether a firm should do as much of both activities as possible OR strike a balance between them: A balanced view assumes that the two sets of activities as two ends of a single continuum, competing for the same/limited resources. To manage the trade-off, firms need to balance Er and Ei. An integrated view assumes Er and Ei are two facets of learning activities, suggesting that firms need to blend Er and Ei. There is a potential complementary effect between Er and Ei. In the context of achieving internal ambidexterity, while He and Wong (2004) show that both dimensions of ambi promote SME performance (sales growth), Cao et al (2009) report that resource constrained SMEs benefit the most from achieving an appropriate balance of Er and Ei.
External ambidexterity on innovation performance In the OI process, we argue that engaging in both inbound and outbound promotes innovation To develop innovative products and services, firms need to have both technological and market knowledge (Dougherty, 1992); Yet, SMEs tend to lack sufficient market knowledge due to limited resources. By carrying out outbound activities, SMEs increase the stock of market knowledge, which helps to identify technological knowledge needs in the future and manage inbound activities more efficiently (learning effect). Provide financial resources for internal innovation development. At the same time, OI activities may have a diminishing return on firm innovation performance (Laursen & Salter, 2006), so we posit: H2a: Firms that engage in both inbound and outbound activities will achieve a superior innovation performance. H2b: Firms that achieve balance between inbound and outbound activities will achieve a superior innovation performance.
External ambidexterity on sales growth In the OI process, we argue that doing both inbound and outbound promotes sales growth A two stage model of transferring knowledge and technology externally 1) to ascertain what knowledge to transfer externally vs. internally; 2) facilitate the transfer - desorptive capacity. So tech k. gained from inbound activities is also required in outbound activities. Inbound activities will also increase the potential pool of ideas for future technology transfer in their own market and others market. Inbound activities also enhance a firm s reputation as an innovation in the market helping attract future partners for external transfer. Firms over-committing to exploitation tend to fall into the trap of myopia of learning (Levinthal & March, 1993), and hence are unable to sustain their advantage in the market. H3a: Firms that engage in both inbound and outbound activities will achieve superior sales growth. H3b: Firms that achieve a balance between inbound and outbound activities will achieve superior sales growth.
The UK~IRC OI survey Sampling frame Drawn from FAME Stratified by size and sector 15 sectors including both manufacturing (65%) & business services (35%). 5-999 employee size. Questionnaire (12 pages) Designed to capture the wide spectrum of firms open innovation practices. We did not use the term open innovation due to both various definitions and ambiguities surrounding managers interpretation of the term. Drawn upon the extant literature. OI experts and pre-test. Between June and Nov 2010, 5 waves sent to over 12000 UK firms, 1202 sample achieved (10% response rate). No sig. non-response bias. Details see Cosh & Zhang (2011) Open innovation choices: what is British enterprise doing? http://www.cbr.cam.ac.uk/pdf/oi_report.pdf Final sample in this study N= 728 (restricted to SMEs with less than 500 employee size) 11
Open Innovation Choices: What is British Enterprise doing? Andy Cosh & Joanne Zhang with Anna Bullock Isobel Milner Centre for Business Research University of Cambridge
Dependent variables: Innovative sales: % of total sales revenue that came from new, or significantly improved, products and services that had been introduced within the last three years. Sales growth: total % change in sales revenue over the last three years.
Independent variables: Inbound OI activities: Breadth: breadth of their firms external search activities the number of different search channels external to the business that they had used over the previous three years: B1=1 if breadth =0 to 4, 0 otherwise B2=1 if breadth =5 to7, 0 otherwise B3=1 if breadth =8 to 10, 0 otherwise Outbound OI activities: Ktrans: whether the firm had transferred its knowledge, or technology, to external parties within the previous three years with the purpose of accelerating its innovation and creating value for the business.
Table 1 Variable definitions and descriptives Variable name Variable Definition N Minimum Maximum Mean Innovative sales % of sales due to new or significantly improved products 671.00 100.00 30.5878 Ln Innovative sales Logarithm of (Innovative sales + 1%) 671.00 4.62 2.5656 Sales growth Total percentage change in sales revenue over last three years? 689-100.00 400.00 24.0980 Breadth Number of external information sources/search channels used 728.00 10.00 6.3915 B1 B 1= 1 if Breadth = 0 to 4, 0 otherwise 728.00 1.00.2212 B2 B2 = 1 if Breadth = 5 to 7, 0 otherwise 728.00 1.00.4245 B3 B3 = 1 if Breadth = 8 to 10, 0 otherwise 728.00 1.00.3544 Ktrans If knowledge / technology transferred externally =1, 0 otherwise 728 0 1.27 BxKy Inbound / outbound combinations x= 1,2,3 and y is Ktrans value Market Largest market in terms of sales revenue (1 if local, 2 if regional, 3 if national and 4 if international) 728 1 4 3.33 Employment No. of Employees 728 1 480 60.30 Ln Employment Logarithm of the number of employees 728.00 6.17 3.4790 Competitors How many firms do you regard as serious competitors? 728 1 1000 12.94 Ln Competitors Logarithm of the number of competitors 728.00 6.91 1.7740 R&D active If engaged in R&D in last year = 1, 0 otherwise 728 0 1.64 Age Firm age 728 1.00 125.00 22.2225 Young Young = 1 if firm age is 0-8yrs, 0 otherwise 728 0 1.23 Teens Teens = 1 if firm age is 9-16yrs, 0 otherwise 728 0 1.23 Midage Midage = 1 if firm age is 17-32yrs, 0 otherwise 728 0 1.35
6 groups of OI firms EXTERNAL TRANSFER No Yes SEARCH BREADTH Low Med High B1K0 N=134 B2K0 N=234 B3K0 N=167 B1K1 N=27 B2K1 N=75 B3K1 N=91 Inbound only N=535 (73.5%) Ambidextrous N=193 (26.5%)
Table 2 Comparison of characteristics of the firms with different open innovation choices Variables Mean N Mean N Mean N Mean N Mean N Mean N age (years) 21.89 134 15.77 26 23.55 234 19.25 75 23.27 167 21.71 91 ln(age) 2.73 134 2.58 26 2.86 234 2.58* 75 2.84 167 2.75 91 young (=1 if 0-8 yrs) 0.28 134 0.19 26 0.19 234 0.35 75 0.20 167 0.27 91 teen (=1 if 9-16 yrs) 0.22 134 0.54 26 0.23 234 0.24 75 0.23 167 0.16 91 midage (=1 if 17-32 yrs) 0.31 134 0.15 26 0.37 234 0.27 75 0.37 167 0.42 91 Employment 50.86 134 46.12 26 57.24 234 51.44 75 69.86 167 72.36 91 ln(employment) 3.16* 134 2.96 26 3.49 234 3.37 75 3.67* 167 3.80* 91 Main sector (=1 for manufacturing, 2 for services) 1.33 134 1.54* 26 1.30 234 1.52* 75 1.21* 167 1.40 91 High tech (=1 for hightech, 2 for conventional) 1.64 134 1.65 26 1.62 234 1.47* 75 1.56 167 1.47* 91 How many firms do you regard as serious competitors? 12.80 134 13.08 26 13.66 234 7.71 75 14.53 167 12.77 91 ln(number of serious competitors) 1.75 134 1.74 26 1.82 234 1.69 75 1.77 167 1.80 91 % engage in R&D 0.43 134 0.58 26 0.57 234 0.81 75 0.73 167 0.88 91 Largest market for sales (=1 for local, 2 for regional, 3 for national, 4 for international B1K0 B1K1 B2K0 B2K1 B3K0 3.12* 134 3.38 26 3.25* 234 3.43 75 3.45* 167 3.49* 91 % of sales due to new or improved products and services 17.47* 122 20.88 25 27.00* 216 47.13* 70 34.02* 155 42.18 82 Total % change in sales revenue over the last 3 yrs 19.83 124 21.29 24 15.15* 228 52.44* 71 23.98 156 32.14 85 B3K1 * indicates that the value for this group is significantly different at the 5% level than all the other firms taken together. This test was applied tolln(age), ln(employment), ln(number of competitors), market, main sector & high-tech.
Table 3 The impact of inbound and outbound activities on innovative sales. Dependent variable:ln Innovative sales model 1 model 2 model 3 model 4 model 5 model 6 model 7 model 8 Open innovation variables B2 0.925*** (0.209) B3 1.076*** (0.223) Ktrans 0.295 0.297 (0.180) (0.181) Breadth 0.579*** (0.135) Breadth 2-0.0355*** (0.0110) B1K0-0.256-0.861*** -1.387*** -1.138*** -1.233*** (0.452) (0.232) (0.308) (0.251) (0.299) B1K1 0.256-0.606-1.132** -0.883** -0.977** (0.452) (0.433) (0.474) (0.442) (0.470) B2K0 0.861*** 0.606-0.526* -0.277-0.372 (0.232) (0.433) (0.275) (0.208) (0.262) B2K1 1.387*** 1.132** 0.526* 0.249 0.154 (0.308) (0.474) (0.275) (0.286) (0.320) B3K0 1.138*** 0.883** 0.277-0.249-0.0948 (0.251) (0.442) (0.208) (0.286) (0.269) B3K1 1.233*** 0.977** 0.372-0.154 0.0948 (0.299) (0.470) (0.262) (0.320) (0.269) Optimal search breadth is 8. Engaging in inbound activities is positively related to firm s innovation performance(h1a supported) Engaging in outbound activities is positively but NOT significantly related to firm s innovation performance. Control variables Ln Employment 0.214*** 0.210*** 0.212*** 0.212*** 0.212*** 0.212*** 0.212*** 0.212*** (0.0719) (0.0722) (0.0723) (0.0723) (0.0723) (0.0723) (0.0723) (0.0723) Market 0.287** 0.260** 0.256** 0.256** 0.256** 0.256** 0.256** 0.256** (0.119) (0.119) (0.119) (0.119) (0.119) (0.119) (0.119) (0.119) Ln Competitors -0.0907-0.0765-0.0709-0.0709-0.0709-0.0709-0.0709-0.0709 (0.0826) (0.0827) (0.0827) (0.0827) (0.0827) (0.0827) (0.0827) (0.0827) R&D active 1.621*** 1.665*** 1.652*** 1.652*** 1.652*** 1.652*** 1.652*** 1.652*** (0.186) (0.187) (0.187) (0.187) (0.187) (0.187) (0.187) (0.187) Young 0.466* 0.452* 0.446* 0.446* 0.446* 0.446* 0.446* 0.446* (0.250) (0.251) (0.252) (0.252) (0.252) (0.252) (0.252) (0.252) Teens 0.0447 0.0536 0.0513 0.0513 0.0513 0.0513 0.0513 0.0513 (0.243) (0.244) (0.245) (0.245) (0.245) (0.245) (0.245) (0.245) Midage -0.249-0.256-0.249-0.249-0.249-0.249-0.249-0.249 (0.218) (0.219) (0.219) (0.219) (0.219) (0.219) (0.219) (0.219) Constant -2.346*** -1.020* -1.011* -0.756-0.150 0.376 0.127 0.222 (0.677) (0.567) (0.571) (0.686) (0.564) (0.609) (0.590) (0.622) Observations 670 670 670 670 670 670 670 670 Log likelihood -1135-1139 -1138-1138 -1138-1138 -1138-1138 DF 24 24 26 26 26 26 26 26 chi2 267.7 260.6 261.9 261.9 261.9 261.9 261.9 261.9 r2_p 0.105 0.103 0.103 0.103 0.103 0.103 0.103 0.103 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Tobit regression was used since the data is right (37 cases) and left censored (169 cases). Fourteen sector dummies were included in each model.
Figure 1 Predicted % changes in innovative sales from inbound and outbound activities B1K1 113%** B2K1 26% 139%*** 53%* B1K0 B2K0 86%*** 28% B3K0-15% 123%*** 37% 9% More inbound B3K1 More outbound More of both (See Table 3) H2a: Firms that engage in both inbound and outbound activities will achieve superior innovation performance. partially supported H2b: Firms that achieve balance between inbound and outbound activities will achieve superior innovation performance. supported.
Table 4 The impact of inbound and outbound activities on sales growth. Dependent variable: Sales growth model 1 model 2 model 3 model 4 model 5 model 6 model 7 model 8 Open innovation variables B2 0.615 (2.933) B3 0.595 (3.173) Ktrans 9.370*** 9.290*** (2.637) (2.632) Breadth 0.664 (1.854) Breadth 2-0.0498 (0.153) B1K0-0.571 0.360-7.780* 2.364-11.60*** (6.500) (3.214) (4.377) (3.539) (4.246) B1K1 0.571 0.931-7.209 2.935-11.03 (6.500) (6.265) (6.874) (6.402) (6.821) B2K0-0.360-0.931-8.140** 2.003-11.96*** (3.214) (6.265) (3.975) (3.010) (3.779) B2K1 7.780* 7.209 8.140** 10.14** -3.817 (4.377) (6.874) (3.975) (4.173) (4.636) B3K0-2.364-2.935-2.003-10.14** -13.96*** (3.539) (6.402) (3.010) (4.173) (3.922) B3K1 11.60*** 11.03 11.96*** 3.817 13.96*** (4.246) (6.821) (3.779) (4.636) (3.922) Engaging in outbound activities is positively and significantly related to sales growth (H1b partially supported). Younger is better and more competition is worse for growth. Control variables Ln Employment 0.704 0.676 0.603 0.603 0.603 0.603 0.603 0.603 (1.022) (1.022) (1.019) (1.019) (1.019) (1.019) (1.019) (1.019) Market -0.750-0.800-0.631-0.631-0.631-0.631-0.631-0.631 (1.662) (1.660) (1.651) (1.651) (1.651) (1.651) (1.651) (1.651) Ln Competitors -2.897** -2.862** -2.894** -2.894** -2.894** -2.894** -2.894** -2.894** (1.192) (1.187) (1.181) (1.181) (1.181) (1.181) (1.181) (1.181) R&D active 2.682 2.758 2.786 2.786 2.786 2.786 2.786 2.786 (2.652) (2.639) (2.625) (2.625) (2.625) (2.625) (2.625) (2.625) Young 13.94*** 13.88*** 13.56*** 13.56*** 13.56*** 13.56*** 13.56*** 13.56*** (3.680) (3.682) (3.660) (3.660) (3.660) (3.660) (3.660) (3.660) Teens 10.22*** 10.21*** 10.66*** 10.66*** 10.66*** 10.66*** 10.66*** 10.66*** (3.590) (3.588) (3.572) (3.572) (3.572) (3.572) (3.572) (3.572) Midage 8.235*** 8.228*** 8.292*** 8.292*** 8.292*** 8.292*** 8.292*** 8.292*** (3.176) (3.179) (3.159) (3.159) (3.159) (3.159) (3.159) (3.159) Constant -6.082-4.455-3.617-3.046-3.977 4.163-5.980 7.980 (9.213) (7.881) (7.876) (9.707) (7.914) (8.572) (8.295) (8.739) Observations 688 688 688 688 688 688 688 688 R-squared 0.086 0.086 0.091 0.091 0.091 0.091 0.091 0.091 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Robust regression was used in order to correct for the impact of outliers. Fourteen sector dummies were included in each model.
Figure 2 Predicted changes to sales growth from inbound and outbound activities B1K1 7.21% B2K1 0.57% 7.78%* 8.14%** B1K0-0.36% B2K0-2% B3K0 3.82% 11.60%*** 11.96%*** 13.96%*** More inbound B3K1 More outbound (See Table 4) More of both H3a: Firms that engage in both inbound and outbound activities will achieve superior sales growth. partially supported H3b: Firms that achieve a balance between inbound and outbound activities will achieve superior sales growth.--supported
Summary of findings Engaging in OI benefits SME performance; and managing to be ambidextrous is better still. Whilst inbound activities promotes innovation, outbound activities promote growth. To promote innovation, there are benefits from achieving an appropriate balance between inbound and outbound activities (B2K1 as the optimal group) you can go too far in inbound activities. To promote growth, there are benefits from engaging in both inbound and outbound but, subject to that, the firm should do as much inbound as possible (B3K1 as the optimal group).
Discussions and contributions On distributed and open innovation literature Provides empirical evidence of openness benefits SMEs. But outbound alone does not necessarily promote innovation among SMEs. So future direction on outbound activities in SMEs. Examining how different forms of openness can be combined. We show that ambi firms achieve superior performance in both innovation and growth (opposing Ebben & Johnson, 2005). But they need to achieve an appropriate (and different) balance to promote innovation and sales growth. On ambidexterity literature A useful addition to the extant literature on internally oriented ambidexterity; pointing to an alternative way of thinking of achieving ambidexterity in SMEs. Suggests there are complementary learning effects between inbound and outbound activities for SMEs. SME managers need to think about how to formulate and implement their open innovation strategy.
Limitations & our future work The role of Internal R&D in moderating the relationship between open innovation strategy and firm performance. Antecedents and drivers of firms open innovation strategy
Activities, forms and distribution of revenue 29% of firms transferred knowledge and technology externally. Both financial and non financial rewards are adopted. Variety of revenue sources. Exhibit 3.2.1 Proportion of firms engaged in outbound transfers, the benefits they receive and the sources of financial gain 70 69 60 50 53 45 40 30 20 29 30 27 19 24 10 0 Transfer of knowledge to external parties For financial payment In exchange for other benefits Free of charge From contract R&D From outlicensing agreements From spinouts From other sources ------------------ Benefit --------------- ---------------------- % of revenue --------------------- Weighted: 57,420; unweighted: 1,095