OPERATIONALIZING NETWORK DYNAMICS
In spite of its popularity, research concerning the dynamics of entrepreneurial networks beyond individual ties is mainly conceptual. A reasonable explanation is that any attempt to furnish network models with data calls for a considerable effort. This challenge for empirical research may be approached with either qualitative or quantitative methods. As in any research into social processes, such as the micro-dynamics of networks, qualitative research has obvious advantages. Interesting qualitative research into such organizing has been provided by, among others, Bouwen & Steyaert, l990, Larson, l992, Frank & Lueger, l995. Curran et al., l993 use a "critical-incident" technique to uncover the role of networking in small-business operations. But these approaches also have drawbacks including limited statistical generalizability and difficulties in researching networks beyond dyads. Quantitative research in turn lacks sensitivity to the details of self-enforcing and trust-building idiosyncratic exchange processes and creates a bias toward structural features of personal networks. These structures represent nevertheless significant frames for/outcomes of network processes.
Two quantitative survey approaches may be adopted to uncover dynamics: the comparative-static design and the panel study. The comparative-static approach means allocating respondents in categories (cohorts) according to stated position or phase in the entrepreneurial process. By sampling ventures in different phases - conceptualization, gestation/start-up, operation - variation in network characteristics according to phase can be assumed to reflect network dynamics. Cf. Greve, l995. The main drawback of this approach, besides its conceptual assumptions and risk for causal fallacy, is that e.g. changes in individual entrepreneurs´ networks, let alone in individual ties, cannot by investigated. An alternative dynamics analysis is the panel study, i.e. the collection of data from the same individuals at different points in time. The panel approach will take care of the major drawback of the comparative-static approach. Instead problems such as panel mortality and a changing overall context over the studied periods have to be dealt with.
An international project (also including Sweden) investigated, adopting a survey design, the way young entrepreneurs manage personal networking in different national contexts, cf. e.g. Aldrich et al., l989. Besides background variables describing the (prospective) entrepreneur and her/his venture, data were collected concerning the general personal network and more in depth with regard to the five people the entrepreneur prefers to talk with about her/his business. The entrepreneur´s direct linkages to these five persons are here addressed as the "primary network", cf. Figure 1. General networking reports classified data concerning (1) the scope of the overall personal network (number of people with whom the respondent has discussed the emerging business over the last 6 months), (2) monthly investment (in hours) in maintaining and enlarging the personal network as well as travel time necessary to make personal encounters possible.
Primary network data include (1) character of relationship (business/social), (2) frequency of exchange (daily, weekly and less frequent), (3) strength of ties between the network mates themselves (well acquainted/acquainted/ strangers/unknown) and (4) the size of each primary network mates´ own personal network (very large/large/small/unknown), i.e. the focused entrepreneur´s secondary network, a part of the extended network (Aldrich & Dubini, l991).
AN EMPIRICAL STUDY OF NETWORK DYNAMICS
The l987 Comparative-Static Study
The original l987 study including 361 Swedish respondents suggests that entrepreneurs who are in the early, i.e. the incubation/gestation, stage of their venturing process invest significantly (test criteria p < 0.10 if not stated otherwise) less time in maintaining and enlarging their existing personal networks than entrepreneurs who already operate a business, cf. also Greve, l995. Established entrepreneurs also spend significantly more travelling time in order to facilitate personal encounters. Would-be entrepreneurs build significantly more socially oriented personal networks with ties which more often originate in social/casual encounters than do established entrepreneurs. This finding is not supported by Greve (l995) in his female-biased sample, possibly because female entrepreneurs generally build more socially oriented personal networks (Johannisson l992). Prospective entrepreneurs also use their primary network more intensively. As proposed above the would-be entrepreneur needs more social and general support to build self-confidence and legitimacy as well as to acquire social resources before being able to - or having to - reciprocate, cf. Starr & MacMillan, l990.
A conclusive finding of the comparative-static study is that social proximity remains important to all entrepreneurs during the researched early stages of their careers. Physical proximity is also important: both prospective and existing entrepreneurs have at least four out of five primary network partners within one hour's drive by car. The two categories of entrepreneurs are similar with respect to strength/weakness of ties in their primary networks or the scope of the indirect, secondary network. The data also reveal that the primary network mates have been known for about a decade.
The Panel Study l987-l993
The l987 network study reported above was replicated in l993 (also in l988). Over the period l987-l993 the business conditions in Sweden changed from very favourable to very harsh. The original panel in l987, thus, consisted of 361 (would-be) entrepreneurs, found within different sampling frames including compiled lists of newly established firms and participants in courses on "How to start your own business". This l987 gross number was in l988 reduced to 265 and in l993 to 158 persons. Panel mortality was analysed using discriminant analysis and t-tests. Considering various background variables and willingness to respond to the network variables, the analyses show that those not responding in l993 are significantly younger, more seldom live in rural area and more often in metropolitan areas than those remaining in the panel (p < 0.05). There are no differences with respect to e.g. sex, education, stage in the venturing process or willingness to respond to questionnaire items concerning personal networking. In addition to panel mortality the internal non-response rate always has to be considered, especially with respect to how it affects reports on network data.
Our sample is dominated by men; in l987 they constituted 73.4% of the panel, in l993 76.6%. Among the l987 respondents, 31.4% had university training, 52.0% a high-school education and the rest, i.e. 16.6%, the compulsory Swedish education (6-9 years depending on age). The male part of the sample is dominated by technicians while the major educational background of the women (would-be) entrepreneurs is business management, probably mainly accounting.
In the panel study the respondents and their ventures were each year trichotomized with respect to phase or position in the venturing process: "no plans" (albeit they had at least joined a course on "How to start your own business"), "planning" (those envisioning a venture or were nascent entrepreneurs at the time of the study), "operating" (those who had already started, inherited or acquired a firm). Over the one-year period l987-l988, 22.0 % advanced or retreated in the entrepreneurial career. Over the six-year period l987-l993, 30.6% were mobile in this respect in their venturing.
Mobility in terms of changed line of business is measured by trichotomizing the operations of the venture into "manufacturing industry" (including producer services), "consumer services and trade" (including e.g. handicraft and health care), and "other" (miscellaneous operations). The findings show that mobility between the two dominant categories - manufacturing industry and consumer services/trade - is about the same, whether we consider the short-term (l987/88) or long-term perspective (l987/93). For example, over the period l987-l993, about every fourth young manufacturing firm turned into a service business, and vice versa.
Another indication of the complexity and dynamics of the venturing process is the number of firms started during the period l987-93: 37 persons or 24.5% launched a new venture between l987 and l993. Out of those already operating a business in l987, 26.7% started another firm during the six-year period. Elsewhere we have suggested that the network may both trigger and facilitate the combination of ventures into a business career (Johannisson, l992).
The variables reflecting "organizing" supportive contexts include location and cooperation with former employer. - Since Sweden is on one hand a sparsely populated country with approx. 20 inhabitants per square kilometre and has on the other hand a very immobile population, the location of a new enterprise is of special interest in most inquiries into new venturing. The regionalized sample distributes as follows: cosmopolitan areas (45.5%), sparsely populated regions (38.2%) and an industrial district (16.2%). Only two (!) respondents have moved at all between l987 and l993. Asking for customer ties to the former employer (yes/no) 24.7% of the new entrepreneurs in l987 had the former employer as a customer, in l988 22.6% and in l993 33.3%. At least in Sweden new venturing obviously can mean both more competition and increased cooperation in terms of extrapreneurship.
Intrinsic Network Dynamics and its Antecedents
Our initial analysis concerns the individual linkages in the primary network. Out of the overall primary ties reported in the l987 study by respondents still panel members in l993, i.e. 398 individual linkages, 103 or 25.9% remain in l993. There are no differences in tie durability with respect to e.g. sex and level of education. The only spectacular finding is that cosmopolitan entrepreneurs are significantly (p<0.05) more faithful than rural entrepreneurs. Primary-network ties should generally be considered to be important since they represent confidants who are assumed to, directly or indirectly, make major contributions to the enactment of the venture as both a concrete economic and an existential project. These generic personal ties supplement, do not replace, elaborate customer linkages such as those to the former employer, above referred to as extrapreneurship.
The primary network and its strong ties may operate both as a shelter for traditional owner-managers and as a take-off for genuine entrepreneurs. Both assumptions are supported by the data. The "survival" rate of primary network ties thus is highest among those new entrepreneurs who are either extraordinarily stability- or change-oriented. Change then is defined using two indicators: (a) a transition in the venturing process from one phase to another and (b) restructured venture operations according to the adopted classification into different industries.
Among lasting network ties, i.e. ties surviving from l987 to l993, there is a change from a balanced social/business orientation (48.0%, business ties thus make 52.0%) l987 towards increased social orientation (60.0%) in l993 (significance: p<0.05). Of the social ties 77.1% remain social over the period while only 55.8% of the business ties remain business oriented. This indicates that many business ties over time, as the entrepreneur enacts her/his venture, turn mainly social. Male entrepreneurs more often than female redefine their surviving ties: more than every fourth has changed from being a business into a social tie over the period. However in cosmopolitan regions the redefinition of business ties is less frequent than in other regions. This suggests that men as well as prospective/young entrepreneurs in rural/small town areas are especially concerned with expanding their genuinely social capital in their venturing career. According to the findings those entrepreneurs who transferred from one sector of industry to another have to socially enforce their networks as well. There are no differences with respect to change in tie social/business orientation between industry sectors as such, nor with respect to level of education or the existence of an extrapreneurial relationship.
Analyzing lasting ties with respect to mate domicile ascertains that the primary network is spatially concentrated at the beginning of the study period and remains so over the six years (82 vs. 78%). A gross analysis in addition reveals that if remaining network partners move, the chances are considerably higher that they move to the same locality as the focal entrepreneur than that they leave it. A careful interpretation - considering the methodology and the small numbers - suggests that the entrepreneur more often attracts members of her/his personal network than is seduced to relocate by her/his network. As stated, Swedish entrepreneurs are quite immobile.
The findings concerning network exchange reported in Table 1 on the next page suggest that over time the portion of the intermediate category - "frequent" exchange - increases significantly. This may be the outcome of a learning process; personal ties should on one hand not be overexploited, on the other be used in order to remain. Yet, individual variations are common. The table reveals that while half the ties keep the same exchange frequencies over the six-year period, a quarter of the relationships show escalated exchange frequencies and the remaining quarter reduced exchange frequencies. Exchange frequency is obviously not an indisputable indicator of strength of tie in personal networks.
Frequency of Social Exchange in Lasting Primary Network Ties 1987-1993
Note: "Infrequent" means less than weekly exchanges, "Frequent" means weekly and "Very frequent" means daily. Due to a very skewed distribution of the l987 responses the variable was recoded in the 1993 study.
An analysis of conditioning factors concerning network exchange frequency shows that women entrepreneurs significantly more often than men increase their exchange activity over the period. Entrepreneurs with a low formal education significantly reduce involvement in exchange as the venture matures. Take-off as an extrapreneur means stabilized exchange frequencies.
Images of Lasting Network Mates´ Secondary Network in 1987 - 1993
|Secondary network||Small||Large||Very large||Total||Number|
Note: The questionnaire included a fourth option for classifying the network mate's own network: "don't know", a category which was negligible in lasting ties.
The findings according to Table 2 suggest quite a turmoil in the perception of the network mate's own network, i.e. the secondary network. Only slightly more than one third (38%) of the images remain the same. In 37% of the cases the perception of the secondary network has been revised downward and thus in one quarter of the cases upward. A possible explanation is that the career as an entrepreneur, among other things, has enhanced the ability to evaluate the secondary network. Another interpretation is that along with the experiential learning as an entrepreneur the own networking becomes more elaborate. Then the indirect resources accessible through the network mates´ own networks are perceived as less needed and therefore less comprehensive.
Controlling the findings concerning secondary networks with respect to background and contextual contingencies reveals few differences. However, new entrepreneurs in metropolitan areas significantly more often than their peers in industrial districts over time include network mates with an assumed large own network. The l993 entrepreneurs who have not starting another business since l987 include in their primary networks mates with a disproportionately large own network, possibly in order to increase their resource potential..
The analysis of the primary network reports data that refer to averages concerning respondents´ reported ties in that network. - The panel data show that the portion of ties in the primary networks which are considered to be mainly social remain stable over the period covered by the panel study - the portion even increases slightly between l987 and l993, from 42.7% to 43.7%. Two counterbalancing forces seem to have created this outcome. On one hand, the business conditions and need for network support vary considerably between the flourishing times in l987 and the unfavourable business conditions in l993. This would indicate an increased need for social support. On the other hand, entrepreneurs who become more established and legitimate would be expected to have access to resourceful business networks and consequently be less dependent upon social support. This was also the outcome of our comparative-static study of prospective and established entrepreneurs according to the l987 data. Cf. however Greve, l995.
Changes in Social/Business Orientation of Personal Networks 1987-1988-1993
|Change in primary network||l987 - l988||l988 - l993||l987 - l993|
|Increased social orientation||40.4% (67)||34.8% (32)||38.2% (34)|
|No change||24.7% (41)||31.5% (29)||24.7% (22)|
|Increased business orientation||34.9% (58)||33.7% (31)||37.1% (33)|
|Total||100.0% (166)||100.0% (92)||100.0% (89)|
Note: To illustrate, 40.4% is that portion of the respondents who in l988 reported a network more socially oriented network than the same in l987.
Table 3 reveals that the stable average data are rather illusive: data concerning individual primary networks show that they actually are quite dynamic both in a short- and in a long-time perspective. This does not automatically mean that new entrepreneurs in their venturing career change their overall personal network. It suggests rather that, depending on what challenges characterize the venturing process, more socially or business oriented ties in the overall network are given priority to. This seems to be the most reasonable explanation considering that individual ties on the average had existed for about a decade already in l987.
The strength of ties with primary network mates interplays with the ties between the network mates themselves. Too strong intermate ties may short-circuit the network at large, too weak may fragment it and reduce its potential as an organizing vehicle in enacting the focal entrepreneur´s venture. Considering the statement that two persons reported to be "well acquainted" as strong tie, the average primary network (disregarding linkages to the entrepreneur her-/himself) in l987 included 29.1% strong ties. In the l993 study the corresponding portion was 28.3%. Also in the other two categories, where network mates are assumed to be "acquainted" or "strangers", few net changes in the panel at large are reported over the period l987-l993. Again, this stability is surprising since the need for a more diverse primary network encompassing a higher portion of weak ties is expected to increase as the venture matures. A more detailed analysis of the dynamics however reveals that the image of the linkages between mates in the primary network has changed considerably within the panel in spite of stable averages. This analysis once again supports the proposition that the part of the network entrepreneurs give priority to varies according to need. Entrepreneurs manage a combination of strong, weak and no-ties, presumably because the need for a tightly/loosely knit network varies according to the challenges to be coped with in the venturing process.
In Table 4 below we report social-exchange frequencies for those respondents reporting data in both l987 and l993. Obviously the network mates are more intensively integrated in the venturing activities in l993 than six years before. As stated, the increased concern for the confidants may be an effect of the less favourable business conditions in l993. Turbulence in individual primary networks is equally high as in acquaintanceship. A possible reason for this variability in the use of the primary network is that it remains very local over the researched period; in terms of averages the portion of primary network mates who live within one hour's drive by car decreases from 83.2% to 77.4% (albeit, again, with considerable individual variation).
Frequency of Social Exchange in the Primary Network 1987-1993
|Total||100.0% (87)||100.1% (87)|
Note: Cf. note Table 1. Sample averages. Level of significance: *** = 0.01
With respect to the "secondary network" the data reveal an extraordinary average stability over the studied period. There is a slight, not statistically significant, decrease in the perception of the network mate's own network: the portion "very large"/"large" is reduced during the period (from 71.7% to 67.5%) while the portion "small" increases equally marginally (from 24.3% to 27.1%). A more detailed analysis again shows that there have been quite dramatic changes in the composition of secondary network resources from the point of view of the individual entrepreneur. A bold interpretation is that, as the new entrepreneur matures as a businessman and develops an own network, the relative (perceived) size and importance of each (primary) network partner's network becomes devalued in the mind of the entrepreneur, cf. also the analysis of lasting ties above.
General network characteristics concern the scope of and time investments in personal networking. Sample averages report almost the same average number of "speaking partners" over the six-month period preceding the survey (data concerning respondents participating in the panel all three years within the parenthesis): 10.8 (11.6) in l987, 11.0 (11.0) in l988 and 10.5 (10.9) in l993. When panel data are analyzed more in detailed, individual variation is, again, considerable, cf. Table 5.
Scope and Dynamics of Overall Personal Networks 1987-1988-1993
|Increased overall network||0.31||0.28||0.24|
|Same overall network||0.44||0.43||0.49|
|Decreased overall network||0.25||0.29||0.27|
|Total (respondents)||1.00 (194)||1.00 (123)||1.00 (116)|
Since time is a scarce resource for most entrepreneurs, (weekly) time investments in network maintenance and renewal are used as indicators of the significance of personal networking to the entrepreneur and her/his venture. In the overall sample the time investment in networking decreases significantly from l987 to l993, from 9.8 to 8.2 hours per week. This time allocated for networking - actually a considerable portion of a normal working week - is almost equally split between network maintenance and network renewal. The fact that time investment remains the same while the activated network shrinks as business conditions become more severe suggests that the personal network becomes more closed when operated in a harsh environment.
Network Dynamics and Performance
Our expectations with respect to correlations between different network variables and their dynamics and perceived success of the venture are for a number of reasons limited. Personal networking has several functions beside improving the financial performance in each venture; these include general motivating, providing social support and building a platform for further projects. Personal networking appears to be a very long-term investment which will not pay back immediately. This suggests that the personal network is a necessary but not sufficient vehicle for success.
It is then not surprising that no distinct patterns appear as the performance criteria are analyzed with networking indicators as independent variables. The findings in the l987 data that a social bias in personal networking does not imply reduced venture and personal financial outcomes is nevertheless amazing. The most spectacular finding is that a large network and time investment in establishing new ties significantly (p<0.01) increase the expectations concerning future growth of the firm. Female network mates also add to optimism. Long-term ties hamper growth expectations but increase the financial outcome (p <0.01). Frequent exchange in the primary network seems to financially benefit the operations and the personal situation. If the primary-network mates are considered to have a very large own personal network the respondent perceives her/his future as an entrepreneur to be brighter (p<0.05), cf. above.
Using changes in the personal-network characteristics between l987 and l993 as the dependent variable explaining venture success, the analysis looses power because of limited number of observations (approx 100). There are however some indications that those respondents who invested more time l993 than in l987 in networking also perceive a payoff in terms of increased venture, and private, financial success. Increasing time spent for establishing new ties feeds optimism into the entrepreneurial career. The data also suggest that those entrepreneurs who have frozen the l987 portion of social ties are better off than the average. Those who have either reduced or increased the portion of strong ties in their primary networks over the period manage financially better in the l990´s than the average. This adds input to the debate concerning the strength or weakness of strong ties in entrepreneurial settings, cf. also Greve l995. Similar findings concern the portion of network mates who are perceived to have a very large own network - all changes in that perception seem to benefit the venturing career. A possible explanation is that some entrepreneurs recognize the mate's own network (increasing portion), some use experienced success to devalue the quality of mate's own network (decreasing portion).
Previous Page | Main Menu | Next Page
© 1997 Babson College All Rights Reserved
Last Updated 1/15/97 by Geoff Goldman & Dennis Valencia
To sign-up for the Center for Entrepreneurial
Studies' publication lists,
please register with the Entrepreneurship WebTeam.