|
by Hudson Barton Summary of Conclusions
Further information is available upon request..... contact Understanding Skype It's the end of the year and time once again to make Skype appraisals and predictions. The Skype ecosystem continues to grow of course, but questions remain. How fast? Is the growth exponential? Is the growth sustainable? When, if ever, will eBay recoup its $2.6 billion investment?Skype provides us with very little information about itself. Once per quarter, we see an accounting of registered users, at last count 135.9 million, a number of limited usefulness in terms of knowing the actual size of Skype's ecosystem. But they also provide two statistical arrays that are updated constantly and which we capture every hour. One is the record of Skype software downloads. The more important and useful record is the concurrent Skype users online. The number appears in the application window of the Skype client, and it is reported every few minutes in an RSS feed from Skype itself. Jean Mercier, "Skype numerologist", started recording it back at Skype's beginning in 2003, so we have a reasonably complete record. As for analysis, Jean Mercier's assistance has been invaluable, but the conclusions are my own. 
In the graph above, we see the growth of the concurrent user statistic. It is a measure of Skype usage, and not of Skype "users", which is a different but related calculation since it depends not only on the number of users but also the time that users spend logged into Skype. As a measurement of Skype's success, usage is at least as relevant as users. In the "usage" statistic, we can see dips during the summer and Christmas holidays, and spurts of accelerated growth at other times. Skype's growth pattern continues to be distinctly exponential, as we will later demonstrate. The Concurrent User Statistic On a daily basis, various factors cause the number of concurrent users to increase and decrease. As people around the world get up in the morning, go to work, leave work, come home and then go to bed, users sign on and off. Various trends are evident in the growth pattern, and by looking closely, we can make assessments of how the Skype system is being used. By parsing the data according to day-of-week and by time-of-day, we can see distinctions in the patterns of usage between geographical regions and between workday and non-workday environments, and between daytime and evening usage. The following graph, borrowed from Nyanyan is annotated to illustrate how daily events worldwide cause Skype usage to rise and fall.
Every 8 hours or so, it is midday in Europe/Africa, or in the Americas (North/South), or in Asia/Pacific. So the statistic for concurrent users at each of those times corresponds to Skype usage in each respective region (with lesser influence from other regions). When one region is in the middle of its workday, another region is typically asleep, and a third region is at home. Differing numbers of Skype users use the service from each region throughout the day. The following table shows the current record high for each region and the times at which they are measured (Greenwich Mean Time - GMT).
The near doubling of Skype usage that this chart illustrates for 2006 is validated by other measurements. For example, in international voice traffic, Skype's share grew from 2.8% to 4.4% in the past year. The actual minutes of voice traffic grew by 6.2 billion minutes, from 7.6 billion minutes in 2005 to 13.8 billion in 2006, a gain of 80% year over year (Source). Regional Analysis: long-term growth is exponential The long-term historical data continues to show that Skype is growing exponentially, with gains each year exceeding the gains of the prior year. In 2006, the growth per month was nearly 50% greater than in 2005.
This is not the first study that has attempted a regional breakdown of Skype usage. In April 2005, Skype released figures of registered names by country, and in October 2005, Ruell Consulting released a study of names (from Skype's "active" directory). The methodologies for these studies are dissimilar in most ways from each other and from the methodology described here. There is agreement however that Skype's early strength, 2003-2004, was in Europe (including incidentally the Mid-East & Africa). In 2005 the Americas (North and South) experienced the fastest Skype growth. In 2006, in spite of a generous promotion with free SkypeOUT minutes in the USA and Canada for most of the year, American usage did not outgrow Europe. The promotion appears to have been important as a competitive weapon, preventing small and under-capitalized competitors from making progress, but there is little evidence that it contributed to growth in Skype usage.
In every year so far, Asian/Pacific growth has lagged the rest of the world seriously. However the perennial under-performer may soon start to catch up with Europe/Africa and the Americas. While this study of concurrent usage does not confirm assertions that Asia is an area of strong Skype growth, we should expect this to change as broadband Internet service becomes more available in Asia, as the concentration of personal computers in Asia catches up to European and American standards. In the future, the pattern of growth in Asia will reflect the pattern elsewhere, so users will remain logged in for longer periods, and this will cause the number of concurrent users to rise. Broadband penetration correlates well with GDP per capita. With GDP per capita of about US $40K, the U.S. has 20 broadband subscribers per 100 inhabitants while Greece, with GDP per capita of about US $6K, has about 3 per hundred inhabitants. As a result, Skype growth in countries like China and India, where broadband and personal computer usage is less common, may someday soon cause these regions to outpace regions now dominated by Western Europe and the United States respectively. The regional distribution of concurrent usage will also be affected by the increasingly popular dedicated Skype devices, the so-called "Skype phones" that connect without computers to either WIFI or to "broadband cellular" networks. The ability to use Skype with smaller and less expensive mobile equipment rather than a personal computer enables growth of Skype in countries and with populations with lower GDP per capita. These devices will have a significant impact in developed economies as well, especially as replacements phones for WIFI equipped homes and businesses that increasingly consider Skype a viable alternative to traditional phones connected to switched networks. Because of this potential to replace switched networks, Skype is currently experiencing resistance from state-run monopoly telephone companies in many countries. In every case, the way these countries are enforcing their monopolies is by refusing to allow Skype to interconnect via SkypeIN (allowing a Skype user to receive a call from the traditional PSTN network). As a result, the opportunity for VOIP in general and Skype in particular to become a viable telephone replacement is constrained. At the end of 2006, there are still only 14 countries that Skype serves with SkypeIN numbers, and "local" numbers have limited availability. 2007 growth in countries where SkypeIN is available will be faster than in countries where it is not, and as restrictions fall, the regional balance will shift away from countries with already developed PSTN connections. Current SkypeIN countries (a long way to go):
The Enterprise Factor Skype's dominant growth engine is still the general consumer. Throughout 2004, 2005 and most of 2006, evening (non-enterprise) usage was growing faster than daytime usage in both America and Europe. Other studies have sometimes claimed that enterprise usage is an important and growing segment of its business. That claim seems to be premature unless one defines "enterprise" broadly enough to include home-based sole proprietor businesses. For the sake of this study, an enterprise usage is that which can be presumed exclusively enterprise; online in the middle of a normal workday but not online in the middle of a normal non-workday. As such, Skype enterprise usage comprises not more than about 20% of the total, and the share does not appear to be rising.
Nevertheless, enterprise usage is important and very noticeable when on weekdays Skype users log on at 8-9am local time, and log off at 4-5pm local time. At times, we can even distinguish regional enterprise usage patterns. For example, in the fall of 2006 enterprise usage in daytime usage in America was growing relative to evening usage in Europe, and for most of 2005, American evening usage was growing relative to Asian daytime usage. If we were to look at enterprise "users" rather than "usage", we would probably find that in certain economies such as the USA and Western Europe a very high percentage of users would be found online during both the day and the evening, thereby suggesting perhaps that they are "enterprise users", but that is another topic. VoIP services from large telephone and cable companies are currently thought to have exclusive ability to serve the complex telephony needs of enterprise. As Skype software and hardware becomes increasingly capable of serving enterprise users, business usage will grow substantially. Factors include:
Possibility for acceleration From September through November 2006, Skype growth in terms of concurrent users was especially strong, and established the strong exponential growth pattern for the entire year.
The quickening of growth may have been connected to:
Skype also gives us statistics on software downloads. Recently, the downloads statistic was indicating new users being added at a faster pace than in the past. While it has now slowed to normal, the bump in the graph indicates increased interest in Skype, and that may portend a return to strong growth in Skype usage once the year-end holidays have passed. How will Skype do in 2007? Skype will add a million new concurrent users every 80 days or so. The all time peak at the end of the year will be around 13 million.
How Big is Skype Really? Again, this analysis focuses on "usage" rather than "users". The number of Skype users is a somewhat different calculation. In its last corporate report, Skype claimed to have 136 million "active" registered Skype names (or approximately 170 million currently). The criteria for "active" has never been stated, but it might mean the number of unique users over the course of a month. As is often pointed out, that definition is excessively generous and not very helpful. To calculate the number of "real" users, one arbitrary rule of thumb is to take a percent of the registration number, typically 10%. By that calculation, the "real" Skype user-base is 17 million. Another arbitrary rule of thumb uses a multiple of maximum concurrent users (typically the daily peak between 15 and 17 GMT), and is usually stated at 2x. That formula too makes the "real" Skype user-base 17 million. Other studies are based on sampling, but they cover only certain countries rather than the entire Skype network. The following methodology is somewhat less arbitrary, in part because it uses multi-year, consistently measured data across all regions rather than a snapshot in time or demographic slice. The formula is the sum of the observed peak concurrent users in each of the three 8-hour geographical zones. It assumes that while there is considerable variability among Skype users worldwide, the average Skype user is online 8 hours per day. The result is about 21.3 million "real" Skype users at the end of 2006, and 33.3 million by the end of 2007. In other words, Skype will be adding new "real" customers at a rate of a million per month during 2007.
While this measurement of regional Skype usage and users is reasonably certain, it does not necessarily help us to measure sub-regionally, country-by-country. Sub-regional measurements require sampling studies such as the one recently done by inStat. In the chart below, you will see sub-regional estimates of "real" Skype users for a few of the major sub-regional areas. If a current sampling study is available, our measurement matches it (more or less) by segmenting the raw usage data geographically, and making estimates of users' time online. More sampling studies need to be done, especially with observations of usage patterns in areas of the world where we may presume that users are online for fewer hours per day because bandwidth is either too expensive or insufficient to do anything with Skype except text chat. The following tabulation of "real" users in specific subregions may be surprising.
Notably, the inStat report and most studies like it are only of the USA, and while they generally shows Skype ahead in the VoIP race, they fail to mention that 85% of Skype users are outside of the USA. If the same competitors were listed with their share of VoIP worldwide, Skype's impact would be unrivaled. The Not-So-Lousy Investment by eBay
Revenues and profits are less than anticipated by the investment community (but not necessarily less than anticipated by eBay). EBay's hope for Skype, as suggested by its $2.6 billion investment, remains a promise that awaits fulfillment. In 2007, Skype will further monetize its success in usage growth, possibly in ways that far exceed 2006. Revenue will come from three basic areas; licensing revenue, integration with eCommerce services, and SkypeIN/SkypeOUT charges. Nobody knows how much it all adds up to. On the negative side, fabulous success as a VoIP carrier inevitably means that SkypeOUT and SkypeIN revenue is limited because calls will increasingly be made within the free Skype network. In the near term, there are three obstacles: SkypeIN numbers are generally unavailable in most of the world, integration of Skype, PayPal, and eBay is not at all well developed, and Skype enabled services & hardware devices are still in their infancy. If these situations don't improve, the financial outlook is dim. On the positive side, these situations will change, slowly perhaps, but inevitably. The typical Skype user may one day supply eBay with revenue whenever he buys a licensed phone, subscribes to online services, patronizes Skype-enabled merchants, or uses Paypal transaction services? It could total hundreds of dollars annually rather than the current $11.xx. With 35 million users by the end of 2007, there is potential for much higher revenue than the projection above. As for competitors, they appear to have not yet figured out how to match Skype's results. Jupiter Research reports that 6.5 million people in the USA today use VoIP, compared to Skype's 2.5 million. 1.5 years ago, there were reports that Skype traffic represented 35.8% of the total VoIP traffic, a figure that could be higher today.(Source). In the VoIP graveyard, AOL has shuttered its TotalTalk. Rumors abound that Yahoo may be in the process of terminating its Voice component, Microsoft is losing IM customers to Skype (according to Jeff Pulver), and Vonage, which at one time was the largest VoIP telephony provider, is tanking along with Vonage stock. There are more than 900 known VoIP providers around the world, but only a very few have any customers to speak of. 86 VoIP providers have in fact shut down in the past 16 months. Others will meet with a similar fate soon (Source). It is hard to imagine how more than a handful of them can mount a serious near-term threat to Skype. Yahoo and MSN, both of which are client-based VoIP services like Skype, have potential but are longshots for success. Another client-based survivor that Skype is surely watching is Googletalk, with Libjingle, which if it ever becomes a true VoIP service in 2007, may yet pose a challenge. Many VoIP carriers had clearly hoped to gain a secure foothold during 2006 and didn't, thanks in part to Skype's free SkypeOUT promotion in America. In the near term, facilities based VoIP services such as Time Warner, Comcast, Cablevision, and Cox are doing well, but in the long run a client based VoIP technology like Skype has cost advantages, flexibility, and worldwide economies of scale that they cannot match. A tipping point may arrive someday soon when the perception of VoIP will become synonymous with the largest carrier. The closer we get to that point, the more difficult it is for a smaller carrier to challenge the leader (Skype). Free communication within the Skype network becomes a valuable feature when everyone you want to talk to is available without an outside connection. SIP (Session Initiated Protocol) based systems of course would be the ultimate free communication network because of presumed interoperability, but it hasn't happened yet. For a long time, interoperability between SIP carriers is likely to be limited to a small set of features, and it certainly won't be seamless. Moreover, the economic model for interconnected SIP (other than perhaps the Google Libjingle model) is incomprehensible. While other carriers add basic VoIP features, Skype is becoming a sort of social network, supporting concepts of group-ware and Web 2.0. Its product is unique, so far, and the network has grown large enough to make these concepts practical in both social and business settings. Relevant features include:
|
