Lies, Damned Lies, and Unique Visitors

It was either American humorist Mark Twain or British politician Benjamin Disraeli who said there are three kinds of lies: Lies, damned lies and statistics. Considering that both gentlemen lived a century ago, it is amazing how applicable those words are today to the way industry analysts measure the success of e-commerce Web sites.

This week, USA Today carried a story about the complaints of some dot-coms that the visitors to their sites are undercounted by the firms that track Internet surfers.

In some cases, the alleged differences are substantial. Yahoo!, for instance, claims its own tracking shows that it gets 100 million unique users each month, but Media Metrix puts the number closer to 40 million.

Sixty million pairs of eyes every month translate into millions of dollars (US$) worth of potential advertising revenues.

Also, strong visitor numbers can help to seal a venture capital deal for an e-commerce startup. What else can a young firm point to as proof it is building an audience for its site?

Daunting Task

Measuring the number of people who visit any given Web site is in itself a difficult task. For instance, a person who does a search of Yahoo! at work and then another at home, and a third at a friend’s house, is counted as three visitors. But the real problem is measuring Web use in aggregate.

The model currently being used is the same one that is applied to measuring television viewers. A sample is taken — a statistically valid one, to be sure — and the numbers are extrapolated to the total TV-viewing population

The Ratings Game

But no one has ever said that television ratings are perfect. Anyone who has ever mourned the cancellation of a favorite show can relate. And getting credible ratings on the Internet is far more complicated than on TV.

For instance, when I turn on the TV, I simply choose a channel to watch — I don’t have to drill down through layers of programming. But very often, my Internet destination is reached through a Byzantine pathway that takes me through a half-dozen other sites — sometimes more. Hit counters click away as I pass through, but often, I am just passing through.

In fact, one of the frequent complaints of sites that say they are undercounted is that portals they partner with for visibility are receiving credit for their traffic.

Too Big to Measure?

And then there is the sheer scope of the Web. With millions of sites and an estimated 80 million users in the United States alone — not to mention millions of international users confusing the data by obscuring national borders — getting a firm handle on where people come from and where they ultimately go on the Web is a daunting, if not impossible, task.

In fact, a niche e-commerce site might get relatively few visitors, but — precisely because it has a narrow focus — be able to convert a large number of them into paying customers. Strong customer service and good pricing could keep them coming back. That would mean lower customer acquisition costs, and before long, higher profit margins. A success story like that could easily cruise under the radar of the ratings firms.

Numbers Still Count

So what is an investor or advertiser to do with all the statistics? Take them with a grain of salt, for one thing, and dig deep beneath the surface whenever possible. Most companies know — and may be willing to reveal — much more about their visitors than raw numbers.

Without violating customer privacy, companies can reveal how long visitors typically stay, where they come from, where they go next, and how often they interact with the site by filling out questionnaires or e-mailing feedback. Information about the quality of site visits can be far more useful than round figures that reflect sheer volume.

That is not to suggest that the numbers generated by Media Metrix and Nielsen/NetRatings are useless, by any means. They are valuable for tracking trends — for revealing spikes in traffic during a holiday season, for example, or after a major burst of advertising. Such data can help outsiders evaluate a company’s marketing effectiveness and make some projections about its future profitability.

But placing too much importance on page views and unique visitors is folly. Some analysts believe an over-reliance on those types of statistics contributed to flooding the e-commerce world with businesses destined to failure.

Dizzying Effect

Of course, it is human nature to be optimistic, and for backers to hold out hope — especially when millions of investment dollars are at stake — that a high percentage of those 15 million (give or take a few million) unique users each month will convert into a loyal customer base.

It is an understandable, but unrealistic expectation. At some point, the focus has to shift from quantity toward quality. What are customers getting for their money? How likely are they to come back? How many of them will preach the site’s virtues to their friends? Those questions are not answered by looking at page hits alone.

When qualitative information about e-commerce sites becomes more widely available and the ratings are finally seen for what they are — a piece of the puzzle, rather than the whole pie — the statistics will become less important, and more valuable.