The network externalities hypothesis is tested in two software market segments, one dominated by priced products (Windows web servers) and the other dominated by free products (UNIX web servers). We examined data samples of commercial Web servers from August 1995 to February 1997. The results suggest that network externalities were present in markets where priced products dominate, however markets where free products dominate did not exhibit network externalities.
It has been widely suggested that information technology (IT) products and computer software in particular are subject to network externalities (e.g. Farrell and Saloner, 1985). When network externalities are present, the value that a consumer derives from a good or service increases as the network, or installed base, expands. Network externalities are likely to be strong in IT markets since most IT products or services support or facilitate some sort of information exchange. Thus, a large and growing network is often perceived to be more valuable (and hence consumers are willing to pay more to participate in it) since it contains more potential exchange partners. Such a large network (or one that is generally believed to become dominant) is also likely to attract additional, indirect benefits, such as more add-on products, skilled workers, books and manuals, etc. (Shurmer, 1993).
Pioneering work in this area has demonstrated the importance of standards (Gandal, 1994) and market share (Brynjolfsson and Kemerer, 1996) in generating network externalities in packaged software markets (single-product networks). This paper extends the existing network externalities literature in two important ways. First, it provides an empirical investigation of a composite goods network, that is a network where two separate and distinct categories of products are required in order for the user to derive utility. Many types of IT can be classified as composite goods including client/server networks, hardware software networks, and content creator/content viewer networks. Second, the study provides for a contrasting investigation of markets where priced products hold the dominant market position as compared with markets where free products hold the dominant market position. There are a number of cases in IT where free products are widely available. In some cases, firms may offer free products in order to dominate related markets. In other cases, user consortia may create a free alternative to commercial products (e.g. Apache web server, Linux operating system).
The context of this study is the market
for WWW server software. This market is well suited for examining network
externalities in composite goods markets since two products are necessary
for a user to derive utility – a client
and a server. WWW software is also well suited for comparing markets dominated
by priced vs. free products, given that during the time of this study,
the Windows server market was dominated by priced servers and the UNIX
market was dominated by free servers.
The base model for this research is derived from Brynjolfsson and Kemerer (1996). This model uses price as a proxy for the value that the average consumer places on a product. The model is extended here to account for the uniqueness of network externalities among composite goods. Hedonic pricing, expressed as a multiple regression model, is used to test for factors influencing the natural log of the server price. The research model for testing is expressed as:
Two data sets were gathered for this study, each consisting of a time series of 19 panels from August 1995 through February 1997. Observations in these two data sets reflect non-shareware, commercial web server products available for the Windows and UNIX software platforms, respectively. In keeping with earlier studies and the restriction of the hedonic pricing method, only priced products were examined. The data yielded 321 observations of Windows products and 165 observations of UNIX products. Observations for this study were obtained from public data, industry research reports, a content analysis of the press, and feedback from vendors. Multiple sources of data were identified and used to verify quality and constancy.
Three sets of results are presented in Table 1. These models were arrived at by examining all possible variable combinations, removing variables insignificant below the 90% confidence interval. The first set of results, presented in (a), reflects a refined model of the Windows web server market that considers the server product’s share, but not the browser share. These results use a model similar to that employed by Brynjolfsson and Kemerer (1996). Model (b) extends the model in (a) to account for the uniqueness of composite goods markets by including the variable for browser market share (BSHARE). The results in (c) represent significant variables when an analysis is run for the UNIX market dataset. Multicollinearity was tested for in each model using both the VIF and Belsley-Kuh-Welsch diagnostics (Belsley, Kuh, Welsch, 1980), and in each model these results indicated that the independent variables were not significantly confounded with each other.
Market Dominated by Priced Goods (Windows Market)
Model (a) offers strong support for the network externalities hypothesis. The value for server product market share (SSHARE) is significantly related to the log of the product’s price. A rough interpretation of this suggests that with all other factors held constant, a one percent increase in the browser's installed base warrants a price that is slightly over 12 percent higher than its competitors (e0.1153 = 1.122; or 12.2%). The degree of this association is likely due to the fragmented server market. The positive coefficients of two of the three standards variables (DBLINK and SSL) also suggest qualified support for the existence of network externalities,. The variable for SHTTP was not significant at the threshold level. SSL and S-HTTP are competing, although not mutually exclusive security standards. The significant premium being placed on the more popular security standard (SSL mean=.3988 vs. S-HTTP mean=.1776) supports the existence of standards-based network externalities.
The proxy for trialability (TRIAL) proved highly significant and yielded the largest coefficient. Firms that offer a trial version were (ceteris paribus) able to price their products nearly 1½ times higher than were firms that did not offer a trial version. The Web allows the firm to reap advantages from distributing demonstration versions at a low, fixed cost (e.g. Nejmeh, 1994). These results also support the notion that users recognize that there are risks associated with adoption (e.g. Rogers 1983) and that trial versions of products help reduce this risk.
The negative DURATION coefficient may be associated with lower average prices as the volume of consumers grow, as the fixed cost of the initial software development is covered by early sales, and as more competitors lower the overall average price of server products. However, the significance and positive coefficient of the TIME variable unexpectedly suggests that server prices have increased over time, all else being equal. These results are contrary to the quality-adjusted price declines over time identified in other studies of IT (e.g. Chow, 1967, Gordon, 1993). One explanation for this effect is that as the Internet continued to grow, Internet software was recognized as being more valuable to an organization and new manufacturers coming to market reflected this perceived higher valuation in higher product prices (ceteris paribus). In any case, the impact of TIME is negligible. Further tests (available upon request) demonstrated that TIME could be removed without significantly altering the results or their interpretation.
The results in (b) reflect a regression performed to investigate if firms that vertically integrated and provided both components of a composite mix (in this case servers and browsers) realized any price premiums vis-à-vis their non-integrated competitors (firms that only provided servers). SSHARE was insignificant in a model containing both SSHARE and BSHARE (not shown). hence model (b) does not contain SSHARE. A positive benefit between product price and browser share is evident in model (b). A rough interpretation suggests that ceteris paribus, a 1 percent increase in the browser's installed base warrants a price that is slightly over 1 percent higher than its competitors (e0.0117 = 1.0117; or 1.17%). Results among the other variables are quite similar in terms of significance, coefficient size, and direction of influence. The t-statistic for browser share in model (b) suggests a higher degree of probability of significance vs. server share (t = 7.6 vs. t = 6.9) and model (b) contains slightly higher adjusted r-squared (0.775 vs. 0.769) and F statistic values (100.976 vs. 97.592). A comparison of these results is interesting in that network externalities related to dominance in the browser market (BSHARE) offer a greater impact than dominance in the server market (SSHARE). However, caution in the interpretation of these results is warranted, as the separate models may not adequately express the relationship between the BSHARE and SSHARE variables.
Market Dominated by Free Goods (UNIX Market)
The UNIX market for Web servers is dominated by freeware producers, which make up over 50% of the market's installed base. This competitive dynamic violates one of the assumptions of hedonic pricing - that suppliers price products as a reflection of consumer’s willingness to pay. All previous empirical studies of network externalities in IT have examined contexts where priced products dominate. However, as noted earlier this condition can not be assumed in all contexts. The lack of prior empirical work under such conditions raises the following question - can a firm competing in a market dominated by free products still capture and benefit from network externalities? While one context is not likely to provide results which can be generalized in all cases, by beginning to look at such questions we develop a frame of reference for informing managerial decisions and for understanding the robustness and applicability of network externalities theory.
In order to test the hypothesis that network externalities remain significant, even in markets where free products dominate, the UNIX market is examined as the representative sample. The sample of 165 observations of priced, commercial UNIX products was examined to identify the refined UNIX model (c). With an adjusted r-squared value of .724, the UNIX refined model (c) yields explanatory power similar to the Windows refined models. However, the results from the UNIX market vary significantly from those of the Windows market. Most notably, the network externalities hypothesis fails to be confirmed in this market. The variable for server market share (SSHARE) is significant, but it is negatively associated with price. Since only commercial firms that charge for Web servers are included in the sample, this suggests that the UNIX market does not show demonstrable network benefits. One interpretation of the significant negative coefficient is that in the absence of rational pricing, network benefits fail and traditional economics take over. Firms that capture larger market share leverage their scale economies and lower their price to try to make their products seem as attractive alternatives to free competitors.
The significant feature variables (SCRIPT, SEARCH, and UDIR) all yield positive coefficients as would be expected, lending credibility to the overall estimates. The ability to link to external databases is the only evidence of network externalities in the UNIX market and this measure is admittedly confounded with quality benefits. Neither of the other two security standards variables was significant.
The time trend, although small, is highly negatively significant. Such negative time trends have been noted in other studies of information technology (e.g. Chow, 1967; Gordon, 1993). This is to be expected in the context studied given that, over time, it should become increasingly apparent to commercial participants that the free products are eroding the market share of priced products.
When data samples of commercial Web servers from August 1995 to February 1997 were examined, the following results emerged: 1) The network externalities hypothesis was supported in the Windows web server market; 2) With appropriate caution noted, the results from the Windows server market suggest significant price benefits accrued to firms that also offered a browser component. The model featuring browser share also showed slightly more explanatory power than did the model that included server market share; 3) The network externalities hypothesis was not supported in a market dominated by free products (the UNIX market); 4) In both markets, firms that offered a trial version of their products were able to command a significant price premium over their rivals; and 5) Windows servers seemed to derive network benefits from database and security standards.
These findings lend much needed empirical support for the network externalities hypothesis. This phenomenon was examined within a previously unstudied, yet large and influential market. The results suggest that the structure of the market’s potential network (e.g. composite goods or single good) as well as the pricing policies of the players (priced vs. free products) may have significant influence over the market. The results of this study suggest the wisdom of vertical integration in software industries and may influence practitioners to further explore the possible benefits of strategies that would allow a firm to offer both products in a composite mix, such as mergers and acquisitions or research and development.
For references see:
http://www2.bc.edu/~gallaugh/dsi98ref.htm