Internet Traffic and Portfolio Returns

We examined whether traffic data on sites owned by publicly listed Internet companies provide information about the future of those companies that is useful in portfolio management. The study shows that when Internet companies are classified into portfolios according to above-median and below-median traffic data, the more popular sites provide significantly better stock returns than the less popular sites. These results may be explained by the superior ability of popular sites to attract advertising revenues and extract greater compensation from affiliated sites. They may also indicate investors' perceptions that the more popular sites have greater network externalities (in which the value of being a part of the network increases with the number of members already in the network) and have the ability to generate higher future profits and cash flows. These results carried through to offline companies with Internet sites in 1999 but not in 2000, possibly because the online operations of offline companies were still not a material component of the companies' revenues and cash flows. The primary purpose of this study was to determine whether the market rewards Internet companies that have high traffic relative to companies that have lower traffic. Portfolio managers may use the results of this study to select Internet stocks that are likely to provide superior returns.Prior studies have shown that measures of Internet traffic can be useful beyond historical revenues in the prediction of future revenues of online companies. Prior research has also shown that traffic measures are important variables, in addition to traditional financial measures, for the contemporaneous valuation of Internet companies. No prior studies, however, have examined whether knowledge of the traffic data can help a portfolio manager select companies for a portfolio that will systematically earn superior future returns.We used reported Media Metrix traffic data for 15 months in 1999 and 2000 to classify Internet companies into two groups—those with above-median traffic and those with below-median traffic. The particular traffic measures we used included number of different individuals that visited the site, percentage of total Web users that visited the site, and an intensity measure of the average number of days during a month that a user visited the site. We then analyzed the returns to the portfolios after holding them for a month following portfolio formation.This study shows that the traffic measures do classify Internet companies into those with superior future returns—during both the Internet boom of March 1999 to December 1999 and the period of disillusionment, January 2000 to May 2000, when many Internet companies had negative returns. Examination of the returns of the two groups in the month after Media Metrix's release of the traffic data and portfolio formation indicates that the portfolio of above-median-traffic companies enjoyed significantly higher returns than those with low traffic. The results indicate that a portfolio manager could have obtained superior portfolio returns from Internet stocks during the period studied by buying access to the ranking data and focusing on the companies with the higher traffic measures.An explanation for these results is that high traffic may translate immediately into greater advertising revenues—one of the most important revenue sources for sites that primarily provide content to their users. Furthermore, high traffic may indicate that the size of the potential market for the site's products and services is larger than for low-traffic sites and that the companies have the potential for greater revenues and cash flows. Finally, the site's ability to attract traffic may indicate the ability of the company managers to implement and execute their online strategies. Companies that were able to attract and build traffic in spite of volatile conditions characterizing the study period may be more likely to continue doing so in the future. Traffic measures may thus be complements to, or even substitutes for, the generally accepted but elusive “value drivers” of Internet companies (such as sales growth). Our results also indicate that even companies that do not generate most of their revenues from the Internet outperformed their counterparts when they had higher traffic into their Web sites in 1999 but not in 2000.