Can Audience-Based Buying and Selling Work?

Lately, I’ve been hearing quite a few industry veterans talking about shifting online advertising’s buying model away from the impression-based system that we’ve been using, toward a model that has worked for traditional advertising for decades. Unlike some of the other models I’ve heard about, this one seems to make a bit of sense.

Before we get into a discussion on the audience-based- model’s merits, a little background…

A consensus is building across the industry that our planning model needs to more closely resemble that of traditional advertising before the larger traditional advertisers will fully commit to online media. In traditional advertising, planners tend to pick audiences based on their demographic and psychographic characteristics, and they develop entire campaigns based on those audiences. In formulating those plans, they use audience metrics like gross rating points (GRPs) and cost per point (CPP) to measure their efficiencies in reaching those audiences.

GRPs are a measure of reach to a specific audience. Most planners are familiar with this formula:

    GRPs = reach x frequency

That is, your gross rating points are a function of the percentage of a demographic audience reached by the ads, as well as the number of times people in that audience are reached by the ads.

However, GRPs can be calculated in another way:

    GRPs = gross impressions w audience universe

Or, in other words, GRPs are equal to the number of impressions against an audience, divided by the size of that audience.

CPP is another metric often used. Your CPP is the amount of money needed to achieve one GRP, thus:

    CPP = dollars spent w GRPs

In other words, if you achieve 10 GRPs against a given audience with a magazine insertion that costs $30,000, your CPP for that magazine is $3,000.

A growing contingent of Web buyers and sellers is beginning to think that we should implement this model online. And it wouldn’t take too much to get it to work, either.

All that would be required on the part of publishers would be a deeper understanding of their sites’ audiences. At a minimum, a subscriber study would be needed to classify a site’s audience into basic demographic and psychographic groups. Publishers would have to project their demographics by asking a sample group of users about their sex, age, household income, and education level, among other things. This information, coupled with measures of unique users, could be used to give advertisers an idea of how many people they can reach with a given site. From there, it is easy to calculate GRPs and CPP.

Publishers can improve their efficiency at reaching specific demographic and psychographic audiences by collecting profile information, tying this to a cookie, and using the cookie to target ads. This will make it possible to avoid the waste that occurs when ads are served to people outside a given audience.

While it is technically feasible to move online buying and selling to this audience-based model, I hesitate to recommend that we do so.

We all know that targeting ads in interactive media is easier and more effective than in any other medium. By moving to an audience-based model, many traditional agencies might concentrate only on advertising to the correct demographic and fail to take editorial context and other targeting options into consideration. If an agency can buy a certain number of GRPs against women who are between the ages of 35 and 54 and who are online, will the agency also be motivated to take advantage of other targeting options like content category, connection speed, and domain-level targeting? I’m not so sure.

This is one of those situations that we need to discuss as an industry. We have the capability to make the switch, but just because we can doesn’t necessarily mean we should. Do we move toward an audience-based model because that might generate more revenue for publishers more quickly? Or do we continue with what we’ve been doing, using the impression-based model? We have to weigh the benefits against the costs.

Personally, I think that if we move to an audience-based model, it might sell Web advertising short. We know that while an advertiser like Quaker State Motor Oil might have a demographic sweet spot of men between the ages of 18 and 54, not all of its customers fall into that audience. In fact, Quaker State might be better off advertising, untargeted, on automobile-enthusiast sites than cutting a deal with a major portal to simply reach its demographic sweet spot.

Sure, traditional agencies can take editorial environment into consideration when formulating their media plans, but as many of us know, traditional agencies tend to place a lot of faith in demographics. It would be a shame to see Internet advertising sold short in that way.

Related reading