OPA: Local Sites Must Find the Right Measurement Tool
Surveys or panels? The resulting audience statistics vary widely.
The Online Publishers Association (OPA) has released a white paper making recommendations for local Web site audience measurement. Differences arose between the two primary methodologies, surveys and panels. OPA recommends both, but cautions they must be used individually.
“The objective we had in putting out the paper was to provide a roadmap,” Pam Horan, VP of marketing and membership for the Online Publishers Association, told ClickZ News. “We’re not trying to say one is better than another, but to clarify ways you should look at this data and which markets you want to go after,”
The paper analyzes data from five services. Firms conducting panel research include comScore Media Metrix and Nielsen//NetRatings-MegaView Local. Firms measuring local audience through a combination of online, phone and postal mail surveys include Nielsen//NetRatings @Plan (online); Scarborough Research (phone and mail); and The Media Audit (phone).
When data from the two forms of collection were analyzed, survey-based methodologies, on average, reported 70 percent higher visitor numbers than panel-based research.
An example cited in the study looked at the number of visitors to LATimes.com. Visitor data differed by one million between two services. The Media Audit, which uses phone surveys to report data, projected about 1.5 million visitors to the Southern California news site. ComScore’s Media Metrix panel reported half a million readers.
Despite differences in audience size, the OPA says rankings and other findings within each dataset are consistent. Therefore, use of a single source serves local sites better than a combination of data points from multiple vendors.
“In order to be effective in how you’re using them, work with data from a single vendor,” said Horan. “The relative ranking of sites across vendors is consistent.”
While survey-based reporting may attract publishers due to higher readership reports, OPA says both methods offer advantages. Panel research reports at the household level, but captures readers’ actions on a site in greater detail. Survey-based data reveals more about the individual respondent, though drawbacks are also evident. Reader data from surveys is mostly top-level, and tends to focus on larger or better-known brands. Survey data are refreshed cyclically, while panel research is collected constantly.
“The paper [acts as] a tool for both publishers and marketers to understand how local is being measured, and the things that should be considered when doing so,” said Horan.
