Ad Vogue: Add Vague

In its early days, Internet advertising offered what seemed to be a distinct competitive advantage over its older, offline kin: measurability. Throughout the history of advertising, it has been notoriously difficult to draw a direct line from a given piece of marketing — a TV spot, a magazine ad — to a sale. "The Net, though, is different," ran one reasonable-sounding line of thinking explored in Wired magazine in 1996. "The Net is accountable. It is knowable. It is the highway leading marketers to their holy grail: single-sourcing technology that can definitively tie the information consumers receive to the purchases they make."

At the time, Procter & Gamble had recently provided support for this thesis by deciding to base payments for its online advertising on click-through — the number of times a Web surfer was interested enough in an online ad to give it a click. Suddenly, all those other, unaccountable forms of advertising, not to mention the media companies that depended on them, were seen to be in serious danger.

But measurability has not turned out to be the great secret weapon of Internet advertising. It has turned out to be an albatross. The problem is that while it’s hard to prove that no one is paying attention to a given television commercial, it’s easy to prove that practically nobody is clicks on a given banner ad. This has led to a perception that most online ads don’t work. The upshot, a recent Jupiter Media Metrix report groused, is that "many marketers are still incapable of evaluating [online ad spending] effectively; they focus on quantitative metrics like the cost per click and cost per conversion, not on more qualitative metrics like lifetime customer value and cost per shift in brand perception." A separate Jupiter report argued that "the actual number of customers that Internet advertising generates is often several multiples above what is tracked directly." And so last month, Marketwatch.com announced that it would no longer even report click-through rates to its advertisers unless they specifically requested them. "It's a meaningless measurement," a company spokesman said in an article in The Industry Standard.

Of course, the actions of a Marketwatch.com are less crucial to marketing trends than the actions of a Procter & Gamble. But still. The click-through has been under assault for quite some time, and the question now is whether an online marketing paradigm will be built around some new, better, more precise group of measurements — or around the squishier evidence that advertising has traditionally relied upon.

Tom Sperry, president of the Atlas D.M.T. technology division of the Seattle digital marketing firm Avenue A, suggests that click-through will not go away but will be supplemented by an array of emerging measurements. Some of the new methods are meant to be even more refined, like "view-based conversions." These track, for instance, whether a Web surfer has merely seen — but not necessarily clicked on — an online Eddie Bauer ad before visiting the Eddie Bauer Web site and buying a sweater. (The privacy implications of all that will have to be examined elsewhere.) This sort of thing is what will continue to add "greater granularity," as Mr. Sperry calls it, to the process of figuring out the effectiveness of an online campaign.

On the other hand, there are new efforts to argue the case for online advertising even when its results are not traceable to any sales at all. Recently the Interactive Advertising Bureau, a trade association, released a study conducted by the market research firm Dynamic Logic, along with related research from some of the bureau's member companies. The studies compared the effectiveness of various online ad types, whether plain-old banner ads or new types of larger formats (increased ad size being described as a "creative innovation.")

The conclusion, not so surprising, is that bigger ads seemed more effective. More interesting is how this was demonstrated: not by measuring click-throughs or any similarly "accountable" standard, but through an online survey in which respondents simply offered their impressions. The studies found that "brand awareness," "message association," "purchase intent" and "brand favorability" generally increased among those respondents exposed to online ads. For instance, a bar chart shows purchase intent rising 10 percentage points among respondents who saw an ad four times.

And what does that mean? There is not space here to spell out exactly how those percentages were put together. But it’s fair to say that such data is a lot less coldly indisputable than a measure of the number of people who saw an ad and then bought something. That is not to say that the figures are wrong or that they aren't useful. But there is certainly some wiggle in them.

Robin Webster, president of the Interactive Advertising Bureau, suggests that the click-through and its cousins appealed mostly to the technologists who shaped the World Wide Web in the early going. Their faith in technology's ability to measure the previously unmeasurable, Ms. Webster says, meant that "a lot of false expectations were set up." Sure, it makes sense for retailers to want to track the way interactive ads affect online sales, she said. But she figures it makes sense to get "back to basics, the kind of measures we've always used." The squishy kind, in other words.

Another new study makes direct comparisons of certain kinds of interactive ads and television spots but does so basically on the older medium's terms. Unicast, an online ad firm that champions "superstitial" ads (described as TV-like efforts that play in pop-up windows), worked with Harris Interactive, a polling and market research firm, to conduct a survey measuring traditional concepts like brand recall, positive-purchase intent and the like. "What's important is the message ads deliver," says Richard V. Hopple, president of Unicast. (The study found — again, not so surprising — that superstitials stack up well against TV ads in most of those categories.)

This retreat from the total-accountability ideal does not mean that interactive marketing itself is on the run. Spending for ads online was only about $300 million back in 1996, but it had grown to $5.4 billion by last year. That works out to about 3 percent of the total ad market, according to Jupiter. Rudy Grahn, an analyst in Jupiter's online advertising group, maintains that advertising spending will still grow faster online than elsewhere. And he says that advertisers' hunger for accountability (Procter & Gamble's included) is not going away.

But in the end, that holy grail of perfect measurability seems no closer than ever, which is probably the biggest nonsurprise of all. Consider the Nike logo on Tiger Woods's baseball cap. Does it make you more likely to buy a pair of the company's shoes? If so, would you admit it to a surveyor? Would you admit it to yourself? Would you even know it?

The difficulty in quantifying the effectiveness of a walking, golfing billboard calls to mind the merchant John Wanamaker's famous line: he figured that at least half of his advertising budget was a waste; he just couldn't be sure which half.

One way to find out once and for all whether all those unaccountable forms of advertising really matter would be for some company to pull the plug on every piece of marketing whose effectiveness couldn't be documented, and see if sales held up. That's a field study I'd sure like to see. Any takers?

A similar version of this column appeared in the August 27, 2001 edition of The New York Times.