The controversy over “unique visitors”

Within the measurement industry, this overabundance of information works itself out in periodic disputes over data — disputes over what information is most important, and over best how to define or collect it. The fault line that surfaces most frequently is between panel‐based and server‐side measures. The current agitation for new standards (detailed below) springs in part from a very public disagreement in 2009 over what might be fairly called the atomic particle of online audience measurement, the “unique visitor.” For the first decade of online advertising, the total number of unique visitors to a site was usually defined as the count of unique “cookies,” deduplicated over the period of analysis. This had become the de facto standard since most sites don’t require a log‐in or authentication. But, it is a “technology‐based” rather than a “people‐based” standard. A single user visiting from multiple computers (or deleting cookies from his or her browser) will inflate the count; multiple users sharing a computer will produce undercounts. In 2006 the Web Analytics Association, representing mainly server‐side measurement firms, published a definition of unique visitors that added the option to use “authenticated users” when available. The precise meaning of this new standard was unclear; according to a recent article in Mediapost, “the goal of the standard was to educate the Web analyst to the most commonly used definition and to encourage vendors to openly document any variances from the standard, given that data collection and processing techniques may vary from vendor to vendor.”¹⁰Interactive Advertising Bureau, published a competing definition of “unique users” aimed mainly at panel‐based measurement services such as Nielsen and comScore, but specifying that census‐based tools such as Google Analytics should conform as well. The new guidelines require that the measurer “utilize in its identification and attribution processes underlying data that is, at least in a reasonable proportion, attributed directly to a person.”¹¹heated debate and drew heavy criticism for being overly vague. The IAB’s standard invited a new set of questions: Will sites be required to collect personally identifiable information? What would the privacy implications be for sites adopting this definition? Is this new guideline ultimately even applicable to Web analytics firms, or can it only be met by audience measurement companies? As a result, what had been a fairly straightforward metric — if one whose relevance was sometimes questioned — now has multiple definitions, used in multiple ways, by multiple firms. The episode suggests that online measurement is hamstrung not only by the abundance of data available, but also by the inevitable contentiousness of even well‐intentioned efforts to define standards in a developed industry. The IAB and WAA have been working together to approve (though not to adopt) each other’s definition of unique visitors, but they have yet to reach a consensus.