Engagement Research: Will Online Media Succeed Where Others Haven't

"Everyone's talking about 'return on engagement,' but nobody really knows what that is right now," Sheryl Draizen, SVP and general manager of the Interactive Advertising Bureau(IAB) told Jack MyersMedia Business Report. One analyst believes he has the answer in a formula he's offering.

At this month's emetrics conference of digital media analysts, Web Analytics Demystified founder Eric Peterson presented a formula with eight components (such as the percent of sessions above a certain number of pageviews, or the proportion of people who use "branded" search terms to come into a site) that, when summed, provide one possible "engagement index." The formula and the measures going into it are transparent and readily available, he says, so anyone can use it or some offshoot to figure out what works best for them. Today's common calculation of average pageviews per visit "is a horrible key performance indicator for trying to understand the value of your audience," Peterson told Jack MyersMedia Business Report. "It's too simple. I can get a great deal of information from one or two pageviews." Yahoo! has filed to patent a technique for measuring user engagement that Peterson calls "much simpler" than his formula.

Still, it is going to be awhile before advertisers are ready to spend based on any "engagement index." Executives at major sites say they are still discouraged from using the latest applications in favor of ones that will serve pages that register new ad impressions. Advertisers want an engaged audience but they pay the same amount whether an ad is displayed for one minute or five. It's also not clear that the most engaged user is necessarily the best for a marketer. Peterson notes that users who are less engaged with the content may be more likely to click on an ad. Gartner Inc. research director Bill Gassman told Jack MyersMedia Business Report there might even be "negative engagement," someone who participates fully in a site but does little of commercial value.

Peterson, Gassman and other analysts do believe that engagement standards will be arrived at - even if those standards are different for different types of sites. "If I want to buy an ad on, say, Second Life or a blog environment I want to be able to buy ads in the areas of that website where people are more engaged," Gassman notes. It's when the formulae and standards can determine not just how much a user is involved in content but also the return to the advertiser that engagement will start to take hold.

The idea behind engagement is simple: A user (or viewer or reader) who's wrapped up in the content is more likely to consume it, and, perhaps, any associated advertising, more deeply than a user who's just flitting by. That engaged user may even be more likely to take action on ad. That's the theory, at least. The practice has yet to be proven, and the gray areas leave a lot of room for interpretation. Some websites insist their users are engaged because they spend a long time relative to other content properties. Others tout a high number of pageviews per site visit, or interactions like posting comments, registering for emails, playing games, watching video, purchasing goods and so on. In the television arena, research firm IAG equates advertising and program content recall with engagement. Myers Publishing's Emotional Connections™ research differentiates between engagement with content, attentiveness to advertising and responsiveness to advertising, arguing that the three are not necessarily interrelated.

A lot of the battle over how to measure success on the Web - not to mention mobile devices - springs from interactive media such as AJAX, Flash (Jack MyersMedia Business Report,June 11, 2007), Microsoft's Silverlight and video. Throw in distributed media like widgets (Jack MyersMedia Business Report, July 30, 2007), RSS feeds, social networking and podcasting, and you have a mix that's much more complex, volatile and harder to define than the previous de-facto standards of pageviews and the number of unique visitors per month. Audience ratings services comScore and Nielsen//NetRatings this summer said they were moving toward measuring engagement to rank sites, largely factoring in time spent on site, while simultaneously lowering the relative importance of simple pageviews. Their moves came months after the IAB called for audits of the services' methodologies. Executives of content sites say that by doing statistical analyses of panels of people who have opted in to allow themselves to be measured, Nielsen and comScore greatly undercount users in aggregate, as well as in certain demographic groups. Nielsen and comScore both agreed to participate in an IAB-initiated audit of their methodologies conducted by the Media Rating Council and last month announced they were moving from the "pre-audit" into the full audit phase.

Next month, the IAB will hold its first planned "Audience Measurement Leadership Forum" to explore the issues, and engagement will be on the agenda, Draizen says.

Eric Peterson can be reached via Eric@WebAnalyticsDemystified.com.
Bill Gassman can be reached through the Gartner website at http://www.gartner.com/it/about/contact_gartner.jsp.

Dorian Benkoil, a regular contributor toJack Myers Media Business Report, is a senior consultant for Teeming Media, a digital media business consultancy. He can be reached at Dorian@JackMyers.com.