In my 1997 column, " The Fallacy of Atypical Web Examples ," I explained why five issues that were debated incessantly at that time were not very important for mainstream websites. Let's see how these five stories have played out nine years later.
Story 1: Yahoo! and Advertising
Because Yahoo made money from ads, many analysts in 1997 thought that advertising would provide funding for most websites. Today, Yahoo still makes good money from ads. The most recent analysis is that Yahoo makes 0.4 cents per page view ($4 CPM) on its non-search pages. (The search pages make 5 cents per page view, or twelve times as much as content pages.)
Because Yahoo gets about half a trillion non-search page views per year, tiny CPM equates to a very healthy cash flow.
I admire Yahoo tremendously: great usability and an impressive ability to integrate a broad spectrum of services and outside content under one roof have led to its current half-trillion traffic. I have no doubt that Yahoo will hit the full trillion soon.
Still, my conclusion is the same today as it was in 1997: most websites don't count their traffic in trillions. As a result, they won't be able to create sustainable profits from banner advertising . (The exceptions, of course, are search engines and sites with classified ads.)
Story 2: The Wall St. Journal and Subscriptions
In 1997, The Wall St. Journal already had a good business in online subscription fees. It now has 764,000 paying subscribers at $99/year, meaning that it grosses about $55 million annually (subscribers to the printed paper get the website at half price; I've assumed they account for half of the online subscribers). The New York Times has 177,000 paying subscribers at $50/year, grossing $9 million annually. (For more details, see the Times article -- sold at $4 to non-subscribers .)
Because the Times started later and offers less value-added content exclusively for paying subscribers, it's not surprising that it currently makes less money than the Journal .
In any case, my conclusion now is almost the same as in 1997: The Wall Street Journal can do a viable subscription business because of its unique status, but most websites can't. However, I do think good subscription revenues are now possible for some targeted sites, such as those offering sports videos or (legal) online music.
Story 3: Disney and Storytelling
Many analysts were enamored with Disney as a media firm and believed in its potential to offer rich storytelling on the Web. As in 1997, I still conclude that most websites are better off empowering users with tools . On the Web, non-linear user-controlled experiences are superior to traditional linear, author-controlled stories (which have their place -- but it's mainly in other media ).
Story 4: The WELL and Community
This is one case where I have to somewhat modify my 1997 conclusion. I stand by my comment that elitist sites like The WELL are atypical. In online communities participation inequality certainly continues to hold: most users are still lurkers. Thus, as long as they rely on user postings, most "community" sites won't truly represent their community.
All that said, there's been steady growth in the Web's multi-user aspect, and more and more services benefit from connecting users to each other.
Story 5: Amazon vs. BN
In 1997, the general theory was that Barnes & Noble was in danger of being "Amazoned" and, if that actually happened, it would clearly show that old-world companies couldn't make it on the Web.
In fact, Amazon did win, and has a dominant position as the biggest online bookstore. Nonetheless, bn.com is still going strong and offers a good alternate site. More important, BN's chain of bricks-and-mortar stores has flourished, even though the company was beaten on the Web.
Most important of all: even though Amazon won the bookstore category, there are many other e-commerce categories in which bricks-and-mortar companies now run most of the biggest sites. Wal-Mart of all companies has become one of the most successful sites, and few companies are as physical and non-virtual as Wal-Mart.
What Was Atypical in 1997 Remains Atypical in 2006
So, in 4 of the 5 stories, the same conclusions hold. In the fifth case, even though social software has become more important, community sites still face many of the same problems as in 1997. Most average websites can't run a "community" because of participation inequality, flaming, and many other downsides of multi-user systems.
In conclusion, my main thesis was correct in the case of 1997: the five sites that were discussed so intensely turned out not to be representative role models for the vast majority of websites. You're better off researching your own customers and designing to satisfy their needs than spending your time reading in-depth profiles of atypical websites and trying to emulate their atypical success.