CORPORATE VENTURE CAPITAL: STRATEGIES FOR SUCCESS HOLLISTER B. SYKES New York University
Currently, about 80 major companies have venture capital programs that were started for strategic reasons-to help foster new business development. Yet the results have been mixed. Some companies consider their programs successful. Others have doubts. Others have quit. To explore causes for this disparity of results, survey data were gathered from 31 major corporations through questionnaires and followup interviews. Only strategic investment programs, where the motivating purpose is to assist corporate new business development, were covered. Programs conducted solely for financial return were excluded. Data on two modes of venture capital investment were obtained: venture capital investment (VU) directly in new ventures and investment in venture capital limited partnerships (VCLPs) managed by private venture capital firms. Corporations were asked to rate, on aJive-level scale, the overall contribution (added value) of their programs in meeting their strategic objectives. In the survey, eight factors were probed to determine their possible effect on the strategic value rating. Four of these factors appear to have a significant influence on strategic value: choice of primary strategic objective, type and frequency of communications with the ventures or VCLPs, return on portfolio investment, and mode of investment (VCI vs. VCLP). Objectives that produce a mutually supportive environment, such as formation of corporatelventure business relationships, are more likely to lead to success. Objectives that induce a potential conflict of interest between the corporations and the venture, such as venture acquisition, may lead to a nonproductive environment and failure of the relationship. Modes of communication that involve direct and frequent contact between the corporation and the venture regarding areas of special or mutual interest produce the highest strategic value. Of questionable value are routine reports and attendance at venture board meetings. Reported return on investment for those programs that had been in operation for five or more
Address correspondence to Professor Hollister B. Sykes, Graduate York University, 100 Trinity Place, New York, NY 10006. Journal of Business Venturing 01990
School of Business Administration,
of the Amencas,
H. B. SYKES
years averaged 14 to 15%. Portfolio ROI% was positively related to strategic value in the case of investment in VCLPs, but no significant relationship was found in the case of VCI investments. Comparison of the strategic value of direct vs. VCLP investment showed opinion predominantly in favor of direct investment if only one strategy were chosen. However, as the interviews brought out, the two programs can serve somewhat dtflerent purposes and be complementary to one another. The most effective combination is one in which the VCLP investments provide contacts with the venture capital community and “deal flow” and the direct investments enhance specific business relationships such as marketing or research agreements. The implication of these results for corporate investors is that program strategic success can be significantly enhanced by a proactive approach involving frequent interaction with the ventures or venture capital firms regarding spectfic issues of mutual interest. One of the most effective channels for interaction is through the formation of business relationships (“strategic partnerships” in current parlance). Financial returns will probably be acceptable and additive to any strategic returns.
INTRODUCTION The sole investment objective of the private venture capitalist is return on capital. On the other hand, the primary objective of most corporate venture capital programs is strategic. The impact of possible capital gains on total corporate results is viewed as minor compared with the potential for development of new business. In the 1960s a number of major U.S. corporations began to experiment with venture capital as a supplement to internal new business development activities. Direct venture capital investment (VCI) programs were started at Dow, DuPont, Exxon, Ford, General Electric, Grace, Hercules, Singer, and Union Carbide, among others. The strategic objectives of these programs were to provide a “window” on potential new business growth areas and to provide a source of potential acquisitions for entry into these new areas. The “window” concept appeals because history demonstrates that major new business growth areas have evolved from new products originally developed by innovative small companies. Melberg and Fast (1980) and Klein (1987) recommend venture capital investment as a way to provide early insight into the potential for a new business market or technology. However, venture capital investment as a means to acquire independent new ventures often hasn’t worked out. Hardymon et al. (1983) note that the venture capital opportunities available to a corporation are restricted by a number of factors that diminish the chance of strategic success. One difficulty with the “window” concept noted by Hardymon et al. (1983) and Rind (198 1) is that corporate exposure to a venture’s proprietary technical or marketing information can be a legal problem. As a solution Rind recommended investment in venture capital limited partnerships (VCLPs) in order to insulate the corporation from direct exposure. In a formal research study of corporate venture capital, Siegel et al. (1988) concluded that the best Jinancial performance is obtained when the corporate venture capital group has nearly independent autonomy and a committed source of funds. However, these attributes did not appear to differentiate strategic performance, except possibly in a negative direction when the primary strategic objective was acquisition. Over 100 major U.S. corporations have, at one time or another, tried venture capital investment as an aid to new business development. Since 1980, the number of corporate programs has increased threefold. However, as many as one quarter of the companies who have tried such investing have since stopped, and a number of others have questioned the value of the activity. Those continuing investment are trying new approaches to improve results. A few consider their programs to be very successful.
CORPORATE VENTURE CAPITAL SUCCESS
This range of experience could result from r;lndom distribution, but it seems probable that certain management approaches work better than others. From prior experience and talks with personnel at a number of companies with active venture capital programs, we assembled a list of eight factors, one or more of which were believed to affect program success: 1. Choice of primary strategic objective 2. Type and frequency of contact between the corporation and the ventures or venture capital limited partnerships (VCLPs) 3. Mode of investment (direct VCI vs. investment in one or more VCLPs) 4. Portfolio financial return 5. Corporate venture capital manager experience and compensation 6. Organizational position of the primary corporate contact 7. Source of direct (VU) investment opportunities 8. Number of corporate investors in the same VCLP By use of a questionnaire, these eight factors were probed to determine their possible effect on the perceived strategic value of the venture capital programs. To measure strategic value, the corporate personnel were asked to rate, on a five-level scale, the overall contribution (added value) of their programs toward meeting the corporate new business development (strategic) objective. Based solely on statistical measures, the first four factors had the most significant effect on strategic value. The results related to these factors will be reviewed in some depth, followed by a brief commentary on findings related to the other factors. Results are reported on two levels: statistical analysis of data obtained from questionnaires sent to each corporation and interpretive comments based on interviews with the corporations.
STUDY SCOPE Venture capital investment as used herein is defined as the purchase of non-publicly traded equity in an independently managed start-up or growth company. Therefore, this study does not include internally developed and managed corporate ventures. Also, this study is concerned with only “strategic” venture capital investments-those undertaken for the primary purpose of assisting corporate new business development. Investment programs managed strictly for financial return, such as corporate pension fund investments or the investments of GE’s former GEVENCO affiliate, are not part of this study. Two generic modes of strategic corporate venture capital investment are followed: investments managed through an independent venture capital limited partnership (VCLP) and direct venture capital investment (VU) in individual ventures. Data on both modes were obtained and analyzed in this study and will be referred to by the abbreviations VCI and VCLP.
DATA BASE Eighty-six corporations known to have venture capital investment programs were asked to respond to a questionnaire covering their objectives and investment management practices. Thirty-three provided data for the questionnaire. Two of these responses fell outside the scope of this study, leaving 31 usable for analysis. The data were supplemented through personal or telephone follow-up interviews.
H. B. SYKES
Profile for Companies in the Study VCI
Median per company
Time since first investment, years Number of investments, total Investment rate, $M/year
4 4 2.3
4 6.5 4.9
Of the 3 1 companies, 25 had made investments in venture capital limited partnerships and 26 had made direct investments in individual ventures. Twenty companies had made both kinds of investments. This provided an opportunity to compare the two modes of investment for relative strategic value. The fractional (38%) response rate raises the question of nonrespondent bias. To check this we called all of the nonrespondent companies and determined that 21 (24%) did not have adequate records to respond because of limited programs or because their programs had been abandoned and 1 1 ( 13%) had a corporate policy against replying to such inquiries. Only 21 (24%) would not respond at all. We attribute this to work priorities rather than any reluctance to reveal their program results. Consequently, we estimate that about 60% of those companies that could have provided useful data were sampled and that the data are reasonably representative. Annual revenues for all but two companies in our sample were in excess of $1 billion. All were industrial or communications companies. The median investment profile for those corporations making VCI and VCLP investments is summarized in Table 1.
The major dependent variable in this study is a five-level rating ( - 1, 0, + 1, + 2, + 3) by the corporate respondents which measured their perception of the overall strategic value of the program to their corporation. Only 36 to 40% of the companies rated program value at the highest level of + 2 or + 3. Twenty to 24% said the program was of nil (0) to negative (- 1) overall strategic value. The distribution of the value ratings is depicted in Figure 1. Because the perceived strategic value rating (referred to as “VALU”) is a qualitative and personal judgment, personal bias of the respondents could affect the answers. To test for bias, the VALU data were sorted into two groups. The “a” group included those respondents employed by or directly responsible for the VC program. All other respondents were included in the “b” group. A chi-squared test indicated that correlation of VALU with the group type was not significant (p > .5 for the VCI data and p > .2 for the VCLP data). However, the “a” group comprised 70 to 80% of the respondents. If there was bias among this group, we believe it reasonable to assume that it affected the absolute rather than relative valuations.
Venture capital investment is a strategy that has been recommended (Melberg and Fast 1980; Klein 1987; Rind 1981) to meet various corporate new business development objectives. A list of possible strategic objectives, drawn from our prior discussions with corporate venture
6 4 2 0 -7
Value Rating m
Distribution of value ratings.
capital groups, was provided in the questionnaire. Respondents were asked to rank them in order of their program priority. Table 2 lists the strategic objectives and the mean of the priority rat&Q for each investment mode (VCI vs. VCLP). Identification of new business opportunities (the “window” objective) and development of business relationships ranked at the top of the list. Typical business relationships include agreements under which the corporation markets products developed by the venture or the venture conducts research in its area of specialization for the corporation. Our primary interest was to determine whether choice of primary objective resulted in significant differences in rated strategic value of the overall program. For instance, Hardymon et al. (1983) had questioned the efficacy of venture capital-related acquisitions. To test this, the objective ranked #l by each respondent was compared with their program strategic value ranking (VALU). Table 3 lists the mean of the VALU rankings for each group of companies that cited the same first-priority objective. In the VCI group, the acquisition objective yielded the lowest mean VALU rating. Of the five companies who listed acquisition as their primary objective, the two with the most experience (investing for five or more years) gave their programs the lowest VALU ratings. Their overall program valuations were - 1 and 0, both lower than the five-company mean. In the VCLP group, the acquisition objective also scored a low mean value. To test the
Strategic Objective Rankings Mean priority ranking VCIS
Develop business relationships Find potential acquisitions Learn how to do venture capital Change corporate culture
2.0 2.4 3.3 4.2
2.9 2.7 3.8 4.0 -
Assist spin-outs from the corporation
Identify new opportunities
H. B. SYKES
Kelation of Primary Strategic Objective to VALU Mean
Primary strategic objective VCI group Identify new opportunities Develop business relationships Change corporate culture Find potential acquisitions VCLP group Identify new opportunities Develop business relationships Find potential acquisitions Learn how to do venture capital
6 12 2 4
1.67 1.50 I .oo 0.25
0.82 1.00 0 0.50
.054 .056 base ,041 --_-
ti,f mean VALU
tested by two-sample
f tests of each mean relative
to the “base” mean VALU.
significance of differences in VALU rankings for the objectives, two-sample t tests were performed and are listed in the column headed “TSp”.’ The numerical results were directionally supported by our interviews with the company representatives. Several commented that initiating venture capital investments with the intention of using them as an option for future acquisition induces a negative environment. The better entrepreneurs and venture capitalists don’t want to lose the option of taking the venture public at some point.
VS. STRATEGIC VALUE
Achievement of any of the strtltcgic objectives would, of course, depend on effective communication between the corporate managers and the venture or VCLI’ managers. The questionnaire provided a list of types of contact, drawn from previous interviews, and asked the frequency of each. To compare type and frequency of communication with VALU, the companies were grouped into three samples according to VALU rating. The first sample, the “failures,” includes programs with VALU ratings of - 1 and 0. The second, or middle sample includes all data from programs with a VAIN rating of 1. The third sample, the “successes,” includes programs with VALUs of 2 and 3. The mean frequencies of each type of communication for the ‘“failure” and “success” samples were compared for significance by a two-sample t test. Table 4 summarizes the types of communication examined and the two-sample significance for the “failure”/“success” pair, referred to by the abbreviation “Tsp.” Data on frequency and type of VCI contact showed significant two-sample VALU differences in only two instances: corporate requests for expert advice from the ventures and meetings with the ventures regarding business relationships. However, this significance was not confirmed hy Pearson correlation of the complete spectrum of frequency data because of wide scatter of data in the middle group (VALUs of 1). Routine contacts and commu-
‘Tbc relationship between pairs of con!inuous variables was tested for significance by two-sample 1 tests of means. Because of the small sample six, significance tests were calculated assuming samples with separate variances. All significance figures are reported as one-tailed probabilities that the null hypothesis is true. Twosample probabilities arc indicated bj, the abbreviation “Tsp.”
VCI group Corporate requests for expert advice from a venture Corporate meetings with ventures regarding business relationships Periodic reports by ventures on their activities Investment opportunities submitted by the ventures to the corporation Board meetings attended by corporate representatives Venture requests for expert advice from the corporation VCLP group Corporate requests for meetings with the VCLP managers Investment referrals from VCLPs Corporate/venture contacts initiated by the corporation VCLP requests for advice from the corporation Corporate/venture contacts initiated by the VCLP VCLP routine reports on venture activities VC1.P routine reports on deals seen Corporate requests for investment advice from VCLPS
to VALU “Successes”
“Failures” Type of communication
nication, such as periodic reports by the ventures on their activities and attendance at board meetings, were of intermediate or questionable value by both statistical methods. The VCJ,P data provided a more definitive confirmation of the VCJ results. Direct communications between the corporation and the ventures or VCJ,P regarding items of special interest (such as investment referrals) were of significant value. Periodic routine reports were of questionable value. Moreover, the significance of the VCJ,P two-sample differences shown in ‘J’able 4 was confirmed by Pearson correlations using the complete spectrum of data.
Formation of Business Relationships Communication between the corporation and individual ventures concerning some activity of mutual interest would be expected lo be more meaningful than routine meetings or standardized information exchange. As demonstrated by the communication data, such interaction is likely to result from the formation of business relationships such as research contracts or marketing arrangements. Data on the number of business relationships formed in the course of each program were obtained and compared with the strategic value rating. In Table 5, the corporate programs are grouJ)ed in ranges according to the number of business relationships entered. Mean program VAJ,U ratings are listed for each group.
H. B. SYKES
TABLE 5 Number
VALU vs. Number of Business Relationships Mean VALU
entered VCI group 0 1-2 3-4 4+ VCLP group 0 1-2 3-4 4f
0 1.2 1.4
6 5 7
0.6 0.4 0.5
base ,003 ,001
,003 base -
0.7 1.3 1.5
10 6 2
0.5 0.8 0.7
base ,065 .I95
,065 base -
*Significance of mean VALU differences tested by two-sample
I testsof each mean relative to the “base” mean VALU
For both types of programs, two-sample t tests indicate a significant relationship between the perceived strategic program value and the number of business relationships entered. Pearson correlations of VALU vs. number of business relationships entered confirmed these results (R = .60, p = .OOl for the VCI group; R = .75, p = .OOl for the VCLP group). To test whether time was a factor, Pearson correlations were also run on VALU vs. the number of relationships entered per year. For the VCI group, R = .4 and p = .03; for the VCLP group, R = .8, p = .006. Our interviews supported the statistical findings. Respondents stressed that communication between the venture and those corporate units having a specific strategic interest is necessary for an effective relationship. Also, there were some strong opinions that the most valuable communication was at direct working relationship meetings with individuals in the ventures, rather than at board meetings. A number of company representatives said they would not make an equity investment in a new venture unless there was some kind of business relationship involved. One said, we will “continue to invest only in companies where we have a contractual business relationship,” and “an operating division must sponsor and take responsibility.” A representative of one of the companies expressed doubts that “pure” equity investments (meaning without concurrent business relationships) would yield insights that couldn’t as well be obtained through a business relationship.
Contacts with the Investment Community A good “deal flow” is considered necessary to provide a wide choice for investment. The VCLP data indicated that there was significant value in the investment referrals to the corporation by the VCLP managers. In the VCI survey we collected data on the sources of investment referrals. Venture capitalists were the most frequent sources (an average of 27% vs. the next highest referral source, corporate personnel, at 20%). Contacts with the investment community, primarily venture capitalists, are developed through VCLP investments and coinvestments with the VCLPs in individual ventures. Effective working relationships are built over time, usually by coinvesting. For example, we
CORPORATE VENTURE CAPITAL SUCCESS
found a significant venture capitalists average percentage five years or more, five years.
(Pearson p = .038) relationship between the percentage of referrals from and the number of years the VCI program had been in operation. The of referrals from this source was 42% for those programs with terms of versus an average of 22% for those programs in operation for less than
STRATEGIC VALUE OF VCI VS. VCLP INVESTMENT An obvious issue is whether one mode of investment is more effective than the other. And, is it useful to have both types of programs? During interviews each respondent was asked to compare the relative strategic value of direct investment in ventures vs. investment in limited partnerships. From the 20 companies that had both types of programs, we obtained 17 responses. Of these, 10 favored direct VC investments (VCIs), six were neutral, and one favored VCLP investments. Although qualitative opinion was predominantly in favor of direct investment, given only one choice, the respondents noted that the two programs can serve somewhat different purposes and be complementary to one another. This distinction was most evident from the comments of those individuals in companies that had been active for many years in both types of programs. In their opinion, the most effective strategy is a combination of both modes of investment. The VCLP investments provide contacts with the venture capital community and “deal jlow, ” and the direct investments enhance speciJic business relationships. Investment first in VCLPs provides a useful learning experience, although it should not be the primary objective.
RETURN ON INVESTMENT
VS. STRATEGIC VALUE
Unlike independent venture capital funds, most corporations do not make venture capital investments primarily for investment return. Those few who have done so generally set up the venture capital program as a more autonomous operation, which in several cases (GE, Grace, and Into) has eventually resulted in spin-out of the activity as an independent fund, often including other institutional investors as limited partners. (Data from these companies were eliminated from our data base.) Strategic value ratings were compared with reported ROI% to determine whether there was a significant relationship. A positive relationship to rated VALU was found in the case of the VCLP programs (Pearson R = .53, p = .02), but was less significant for the VCI programs (R = .36, p = .12). These results could mean that higher financial returns increase strategic value or that the strategic value ratings were biased by the level of ROI% achieved. Although return on investment is not a primary corporate objective, it seems logical that if a venture cannot survive commercially, it is a poor prospect as a “window” on opportunity. More critically, it would be a poor candidate for a business relationship. The spokesman for one corporation stated that they have a policy of investing only in those ventures in which venture capitalists have already invested. He considered this an important test of economic viability. Our ROI data showed no correlation with the number of years the VCI or VCLP programs had been in operation, which would indicate no experience effect. However, factors other than experience are involved, for example the methods of evaluating portfolios before there are earnings or public trading of the shares and the time it takes for failures to show up.
H. B. SYKES
In assessing average financial results of the venture capital programs, we exciuded data from those programs that had been in operation for four years or less. The averaged annual RUI for the nine VCI programs that had been in operation for five years or more was 15%. Four reported returns of 10% or less, and four reported returns of more than 20%. The average ROI for 11 VCLP programs was similar, at 14%. Five reported returns of 10% or less, and four reported returns of more than 20%. On balance it can be concluded that strategic investment programs for the past few years have had a positive financial effect. So the strategic benefits, if any, did not come at an average net cost to the companies. OTHER FACTORS The primary source of irzvestmenr opportunities for direct investment was referrals from venture capitalists. As noted previously, the percentage of referrals from venture capitalists increased significantly with the term of the venture program. However, we found no significant co~elation between program strategic value and the dominant source of referrals. Neither the term of the program nor the years of venture capital experience of the corporate program managers showed a significant correlation with strategic value. However, there was a very significant correlation (Pearson p = .006) between the term (years) of the program and the number of corporate venture capital program personnel who left each year to join private venture capital firms. The primary contact person between the corporation and a venture, after the investment has been made, can be from several organizational locations, e.g., the venture capital program group, R&D depa~ment, or line operating divisions. We found no signi~cant strategic value associated with these three locations or the position level of the contact person. However, a s~kesman for one of the highest-mted programs expressed the opinion that, to satisfy the “window” objective, the primary contact should be a high-Ievel person with broad contacts t~ughout the corporation and credibility with both top and operating management. In the last few years “focused” VCLPs, dedicated to serving the specific business area interests of a sole corporate investor, have made an appearance. This is almost like having a direct investment portfolio, except independently managed. Our data were too sparse to compare this mode of investment with investment in the more common VCLPs which have multiple investors (limited partners). Only two companies in our survey were sole investors in at least one VCLP in addition to being one of three or more investors in other VCLPs. Both companies reported that the soie investor relationship was more strategically effective.
Several implications for management can be derived from the results of the statistical analysis supplemented by the interview comments. One failure mode of corporate venture capital may be pursuit of the wrong objective. Objectives that produce a potential conflict of interest between the corporation and the venture can lead to a nonproductive environment and failure of the relationship. Objectives that produce a mutually supportive environment can lead to success. Corporate conflict of interest objectives include: l
A unilateral corporate desire for acquisition
* Expectation of strategic information input in exchange for equity investment only
for the primary purpose
such as marketing agree-
Corporate assistance to the venture in areas of the corporation’s for a window on emerging technologies and markets
expertise in return
The working relationship between the corporation, the venture, and the venture capital community will be improved by emphasis on organizational and management strategies that: l
communicate awareness individuals concerned,
of each other’s specific needs and interests
balance the needs of one party with the motivation party, and
to fill those needs by the other
Corporations should continue to employ venture capital investment as one mode of remaining alert to new opportunities for business development in areas that relate to or could be extensions of their existing business. Use of venture capital to explore entirely new, unrelated business areas also may be of value, but development of effective communication channels and implementation of follow-on strategies will be more difficult because it will be more difficult to find areas for mutually beneficial business relationships.
REFERENCES Hardymon, G. F., DeNino, M. J., and Salter, M. S. May-June 1983. When corporate venture capital doesn’t work. Harvard Business Review 114:120. Klein L. E., Winter 1987. How a venture capital initiative can help the corporate ‘intrapreneur’: A case study. Business Development Review 1(4):22-27. Melberg, R. S., and Fast, N. D. 1980. Identifying new business opportunities. SRI International, Business Intelligence Program. Guidelines No. 1053, November. Rind, K. W. 1981. The role of venture capital in corporate development. Strategic Management Journal 2:169-180. Siegel, R., Siegel, E., and MacMillan, I. C. Summer 1988. Corporate venture capitalists: Autonomy, obstacles, and performance. Journal of Business Venturing 3(3):233-247.