|
|
Digital
Kapital Part 2 on assessing Internet opportunity Review: The story thus far - in past issues we've discovered brand equity is a foundation for predicting the future traffic on a site, but also that it is just one aspect of the calculation. You must also include estimations of the cost of supporting the brand, driving traffic, enhancing technology and, beneath it all, creating interesting stuff, like editorial content, chat environments, and so on. The bulk of Web value has come straight out of the human mind and been swept into the economic sphere by Web technology. What was avocation is now potential vocation, as millions of enthusiasts lay down layers of informational value that can be mined by Web entrepreneurs. The question for this issue - what are the back-of-the-envelope calculations to use in assessing the value of a site? EVERY site has its price, in terms of the time it took to build it, the cost of construction, or the expense of advertising on it. So, one day you're sitting around your swank offices and the phone rings. An eager voice tells you they have a business plan for their site, that they got your name from a friend of theirs, and they'd like to raise money; perhaps they'd just like to sell it to you, because it's such a great idea. The caller has their idea of what the site is worth, not to mention what it can become. If we're looking for gold deposits in the Internet, we need to have the same kind of mental cues that a prospector in pyrite country used to pick out a good site to pan for real gold. This is, after all, a World Wide Web with 2,308,502 sites, according to the May Netcraft survey. There are about as many nooks and crannies where a good business might hide as there are creeks in the American West. Finding gold is no simple job. For instance, I recently received several mailings about a site for sale. They include claims that the site has an average of 70,000 visitors daily, and more than 6.09 million pages served in three months. Taken in isolation, those seem like really big numbers - just what do they mean, though? Besides the fact that they seem to be figures made up by an unmathematical mind? A little math shows that 6.09 million pages served to 70,000 daily visitors is less than one page per visitor per day - that's not much of a business. A later claim by the would-be seller, that the site attracts 250 million viewers annually, is a 10x exaggeration (70,000 visits daily times 365 = 25.55 million visits). If, as I've written in previous issues, value is something that accretes to sources of context, the current craze for "portal" businesses makes sense. This strategy, though, just plain misses the majority of real value on the Net. Netscape, Microsoft, CNET, Disney, Excite, Lycos, and others, racing to position themselves against the only two de facto portals, Yahoo! and AOL, sense that providing context is valuable. The problem is that context is a narrow field surrounding each individual or topic. The portal strategy will work for a few sites certainly, but once the consumer dives below the skin of the Web - a field that provides context about the Web as a whole, not a particular topic - there are myriad vertical sites to be built at a profit. So, when someone says, "My site's a winner," what are the questions that you should ask to determine whether they are full of themselves, or just full of it? This is the first step in choosing to invest or to do a story for a newspaper about the local Web phenom. Performing these calculations demands that you keep up on the basic demographics of the Web. As mentioned, there are about 2.3 million Web sites and somewhere between 44 million and 53 million World Wide Web users, depending on which poll you want to endorse. When you add email users to the Web population, you get approximately 110 million Internet users worldwide. No one keeps track of the number of pages served across the Web, the preferred metric being the number of visitors. However, because the number of visitors means all sort of different things, from unique IP address to sessions, there's no standard mode. After all, dial-up Net users are assigned IP addresses dynamically; each time they log onto the Net they have a different IP address. A Web audience is like a muddle of ghosts being blown through the wires connecting all the computers in the world. From one moment to the next, the address space describes a different set of people. I think we're stuck with the number of pages as the firmest measure of a successful site, at least for back-of-the-envelope math. Number of pages is related to number of advertisements a site can serve. Pages-per-session correlate with informational value, since they serve as an indication of the visitor's involvement with the site. The point is to arrive at a sense of the traffic per client computer and, by extrapolation, what the threshold number of pages served represents a successful site. And, as I said, no one knows exactly how many pages are delivered to Web users each month. Which leaves us to determine the typical Web user's page consumption. If I'm typical, and I am not, it's in many thousands each month. Fortunately, most people have better things to do with their time. Watching non-computer professionals and young people on the Web, the typical monthly page count is between 300 and 500. Then, there are a lot of folks who barely use the Web, surfing to 50 or 60 pages a month; many less. So, for simplicity's sake, let's say the mean Web user's page consumption is 10 per day. Shortly you'll see that this produces page view numbers that are close to reality. So, you've got 45 million Web users looking at 10 pages per day, as an average. That means that each month 13,687,500,000 pages are dished up over the Web . If you take Mark Lotter's Internet host survey, which shows how many computers (clients, too, not just Web servers) were connected to the Net during his monthly poll, and divide the total number of pages served by the Lotter result (29.7 million in May, 1998), you get an mean of 461.32 pages per host each month. Keep in mind, some of these hosts connect many thousands of computers to the network. The result is in the range of pages I suggested each "typical" user downloads over the course of a month, so let's assume that the 10-page-per-user figure is a good basis for doing the rest of this cipher. Next, we have to invoke Pareto's Principle, concocted by Italian proto-Fascist Vilfredo Pareto, who postulated that 80 percent of any outcome was produced by 20 percent of the input. Pareto was an economist, but his analytic strategy, known as the 80/20 Rule to managers today, is applied effectively in all sorts of fields. It's one of those postulates that chaos researchers will figure out someday. In other words, 80 percent of Web pages are served by only 20 percent of the servers. This, though, is a limiting concept. I argue that, for purposes of assessing value, you have to apply a cascading Pareto analysis, finding not just the top 20 percent of sites, but also the top 20 percent of the top 20 percent, and so on, for several iterations of the calculation. To find a median number of pages served to identify a potentially valuable site, I've also opted to divide the number of Web sites in the top 20 percent by the total number of Web pages served each month. And at each iteration of the calculation, each time we apply the 80/20 Rule, I'll divide that number by the total pages served. This is a shorthand trick to correct for the massive volume of Web spider traffic reported as authentic traffic by many sites. The preparatory calculations, then, are:
Remember that the number of sites changes constantly. Here's what we're talking about, in hard numbers during May, 1998:
Investing in a Top 20-percent site is no guarantee of success. Sites lying in the top fifth among Web traffic are merely interesting, and may demand a second look. I'll guarantee you that there aren't 92,340 Web sites out there turning a solid profit. The real value lies in the top 0.8 percent sites on the Web, a select 18,468 sites in May, 1998. When it comes down to it, the real winners in today's market already live in the top 20 percent of the last of our groups, "a," a class of double-a ("aa") sites among the top 0.03 percent on the Web. This group accounts for about 700 sites that deliver a mean of 19.5 million pages per month. The real trick is to pick sites from among the top 0.8 percent, our prime 18,468, in May 1998, that can leap into the very top tier of "aa" sites. All we've done, so far, is narrow the field of possibilities to candidates that are interesting or very interesting. What are the characteristics of those sites. For now, the criteria we will use is strictly numeric, the intangibles will be the focus of future installments. We now have to divide the total number of pages served on the Net each month by each of the Top 20 variables, n, y, z, and a to find the mean number of pages served by sites in each strata of the cascading Pareto Analysis. To calculate the total pages per month, multiply the number of Web users times 10, take the result and multiply that by 365, then divide by 12. In May, 1998, the mean number of pages served in each strata of the Top 20s are:
Again, let me stress that these are merely shorthand guidelines for assessing how much value might be lying hidden in a Web site. There's so much more to consider, and we will in future issues. But, for a first cut mechanism, this is as good a cheat sheet as I've seen. Checking the number of pages served against these rules is akin to looking at a valley, the high walls that surround it, the course of the streams that cut down the sides, to figure out where gold might have collected as it washed downhill. If we're going to find gold, we must begin to pile up these kinds of tracking skills, so we can tell at a glance whether we want to stop, pull out the pan and sift for gold, or go on to the next opportunity beyond the hills ahead.
|
|
DoJ: Present and Future My comment in the last issue that the DoJ action against Microsoft will lead to a slowing of innovation raised a lot of hackles. The point is not that a single event will precipitate the fall of the market. Microsoft is an idolum (Lewis Mumford's coinage for "a symbolic milieu composed of images, sounds, words, fabrications, and even natural objects to which man has attached a representative value") that is an important buttress of the current economic state of the world. People have placed a lot of their hopes on the idea of Microsoft and all those millionaires it has produced. Knocking down that bulwark of optimism about the wealth-creating capacity of the information economy, along with the potential attack on Wintel co-monopolist Intel, at a time when the Asian crisis, full-employment and apparent jitters in the market, is a dangerous game. The DoJ could easily be blamed for "causing" a major correction in the market, making it more of a pariah than it has been portrayed to be by anti-regulation agitators. I take an altogether contrarian view here, since I hold that variations on the themes played by both sides of this debate are true. Apparently, I disagree with almost everyone. I believe the DoJ is wrong to pursue its suit against Microsoft for wrongs it sanctioned with a toothless consent decree in 1994. At the same time, I advocate a proactive injunction against Microsoft's extension of its computer OS monopoly into other media, like digital television and satellite-enabled navigational services. I believe Justice should, at this point, force Microsoft to create Baby Bills that are separately funded to compete for new areas of business. However, the cost to consumers of preventing the availability of an integrated set of features in Windows -- no matter how poorly made Windows is compared to the separately distributed features of third-party ISVs -- is too high. Likewise, because of the potential chilling effect on software investments due to fear of intervention against successful competitors, the DoJ action poses a serious psychological threat to the sustained and exaggerated bull market of the past decade. It could all come crashing down if VC monies are decreased by more than 40 percent in the next two years. New companies and IPOs feed the momentum of the established markets, plain and simple. There's nothing rational about it, as any layman-psychologist can tell you. There needs to be a wall erected around Microsoft, one that encloses the computer industry, as well. The message should be, if you want to benefit from the stability of the Windows operating system marketplace be prepared for competition from Microsoft. If companies want to extend the value of the microprocessor to a new arena, then they should be allowed to enter that market on equal footing with a Baby Bill, should Microsoft wish to compete there, as well. I've been deeply absorbed with Ron Chernow's Titan, a book getting wide coverage right now -- albeit grossly oversimplified coverage in the comparisons to Gates. The shared Messianic qualities of Rockefeller and Gates are obvious, but it is the essentially unformed markets that they entered early and each came to dominate I find interesting. Before Rockefeller and before Gates, there was no coherent path from launch to liquidity in their markets. And neither favors (favored) unrestricted competition; nor do I believe any of their adversaries do, either. In both epochs, it is clear to me that the influence of government can bring a beneficial level of democratic choice to the market. It's also important not to extend the comparison too far; for example, it's not necessary to find analogues in the computer industry for each of John D. Rockefeller's rivals in the oil business (despite the propitious presence of rivals in both industries named "Jim Clark"). Beneath it all, both Rockefeller and Gates, as well as their adversaries, perform calculations of benefit that exclude the consumer. They battle over spoils and the maximization of returns, but the consumer sits by waiting for the benefits to trickle down to them. In Rockefeller's time, he reduced the cost to consumers of oil refining by one-half, but within the industry, he created vast economies of scale that actually reduced the cost of refining by more than ninety percent. He objected to the idea that his margins could be controlled, and sustained the antitrust action long enough to organize the post-Standard companies for excellent returns. The long-term effect of the antitrust action was to maintain the very high returns to investors; consumer prices were, in fact, negotiated and fixed according to the definition of adequate returns. Remember, Rockefeller became much richer as a result of the break-up of Standard Oil, because hidden assets and massive amounts of good will were accounted for in the new companies -- ironic in the wake of Rockefeller's discounting of good will in acquisitions of refineries during the 1870s. Is it bad that the OS and productivity applications markets are relatively impervious to competition? I don't think the OS is impervious to competition, nor do I think it's terribly bad that word processors have stabilized at the point they have -- there's been relatively little improvement in productivity applications since the late 80s. The better-layout applications, like PageMaker and Quark, have never contested the consumer market with simpler and less expensive products, so the consumer has remained in an upgrade-serviced holding pattern. Microsoft is not guilty of this as much as the entire industry. Netscape and Sun have "challenged" the Windows empire with partial solutions to what Windows does. If they'd ever shown up with an adequate competitor to the Microsoft OSes, I'd probably feel a lot different about the DoJ lawsuit. But, really, when it comes down to it, what's the difference between a Windows-dominated OS market and a Java market where Sun controls the Java interface and standards? The industry hasn't challenged Microsoft, because it is complacent with regard to the benefits of a systems platform that provides stability for the exploitation of the consumer. The value of a generic platform, Windows and Intel's x86 architecture, has been at least as important as any other factor to lowering the cost of software development. Face it, developing for Windows gives the ISV access to 90 percent of the computers on desktops around the world. It's sticky to argue about this as de facto co-conspirators in the perpetration of an IT hegemony on the consumer's pocketbook, since everyone in the computer industry is, at heart, as protective of margins as Rockefeller or Gates. We all want to deliver innovation, but only at an adequate price. In essence, the computational benefits of Moore's Law have allowed the IT industry to invite more players to the table to share all the pizza piling up (otherwise, where did all these content players come from?), but the price of a slice to the consumer has stayed relatively stable. The industry delivers more to the consumer (including corporate IT) in exchange for a steadily growing portion of the total economy. None of us are virgins, nor are we saints who have recovered our sinless states after a youthful life of debauchery, Saint Augustine notwithstanding. In order to provide as much (albeit often inferior) functionality as Microsoft does, it is necessary to rob from Peter to pay Paul. When Peter is played by a company, like Netscape, the complaints are audible, because they can afford to hire Bob Dole and Robert Bork to lobby on their behalf. But, when Peter is the consumer, Peter has happily handed over their $99 for upgrades and, except for the occasional bug-inspired complaints, accepted their lack of representation, Jesse Berst notwithstanding. The consequence of the DoJ action, if it is to have any teeth, should be the injection of a consumer advocacy into IT industry decisions. This does not mean that the IT industry should or could submit every decision to the people; rather it means that the people will, at least, have a seat at the negotiations amidst the unfettered competition. The people will figure out what to do with this power. If markets are things that insiders control through legislation of adequate returns, then the consumer should be placed in the legislating body by the government. Information should be made freely available, and the government should act as interpreter to the consumer to ensure intelligent buying decisions are possible. This is the role of the press, as well. But I think the much of the press, both general interest newspapers and magazines and the trades, are de facto insiders representing an equally influential cadre of media companies who benefit from the status quo -- they take and protect their margins, too. The Web-savvy solution would be for the DoJ to negotiate a boilerplate contract for use of Microsoft OS products and make it available on the Internet for anyone to use in negotiation with Microsoft. Make it a "standard" to the degree that, based on widespread knowledge that the terms it lays out are the ones every OS licensee agrees to, any hardware OEM can gain access to Windows without signing a contract that constitutes restraint of trade. Heck, the DoJ should pursue a knowledge-giveaway approach in general, making a public record of the acceptable terms for technology licensing across the board. We cannot endure any more secret agreements between companies, nor between government and companies, as many aspects of the 1994 Microsoft consent decree were concealed from the public. If everyone who went to Redmond or Santa Clara to negotiate with Microsoft or Intel, respectively, knows the competitive landscape, implementation rather than simple access to technology will be the valuable differentiating factor in their product offering. Ultimately, based on the information the DoJ provided, consumers and the press would be in a position to gauge precisely how much value each OEM brought to the market. Imagine the negotiations that would go on in the channel, then. I am not optimistic that the industries and their investors will welcome the consumer to the table. Should it come to pass, we will enter a true one-to-one marketing phase of economic development. Today, the IT industry merely presents different facets of its monolithic structure to the consumer as a simulation of personalized, customized interaction.
The Mocking Corner Businessweek is adding free email services to its Web site. Business Week Online general manager David Smith told CBS MarketWatch: "Every survey tells us that time is the rarest commodity for business people today. By offering e-mail on the Web site, where our readers go for business news every day, we enable busy executives to complete two daily and essential tasks at the same time." Free email a rare commodity? It is attractive to consumers, because it offers them a stable address in Cyberspace. For the businessperson, whose email address must serve a double purpose, providing both a logical location and affiliation information, the idea that they'd get more convenience from an @businessweek.com address is laughable. An extra address adds complexity, rather than reducing it. Now, if that email address were linked to a job search service
that sent "check your @businessweek.com mail" notifications
when a new job opportunity came along to the subscriber's business
email address, then Businessweek would have a credible offering. |
|
The
Library |
|
Sites on my mind: Far
Eastern Economic Review
|
|
|
|||
|
Copyright
1992-2002 by |