By Brad Geving, Head of Media Buying & Ops, Tatari
For ages, ad-buying firms offered one strategy to advertisers for reaching the right viewers: targeting based on predicted demographic, behavioral or psychographic metrics. This strategy has its shortcomings, however, and advances in measurement of actual responses mean advertisers should reevaluate the importance placed on predictive targeting in favor of a measurement-based testing strategy to identify the best targets.
A predictive targeting strategy uses intuition or research to identify the ideal customer and goes all-in placing ads on channels favored by those customers, but the strategy has multiple failure points. The ability to accurately measure response results of ad placements means marketers can instead employ a testing strategy that involves rudimentary assumptions about the ideal customer, placing a small test budget on the ads, then analyzing the results to determine the way forward.
The strategies aren’t mutually exclusive; advertisers would be smart to work with ad buyers that do some of each and make a cost-benefit analysis as to just how much targeting is worthwhile.
On the benefit side of the equation, targeting saves you money, in theory, by serving ads only to those interested in your products; as such, you place ads only where you expect they will perform well. On the cost side of the equation there are fees for purchasing research or data from third parties to make the targeting estimations. In addition, targeting alone will not let you know if you’ve hit the mark and it’s a costly mistake if recurring ads reach the wrong audience unbeknownst to the buyer.
A measurement-based testing approach fills in the gaps by getting to data that’s beyond the reach of a targeting-only methodology. Not only do you gain confirmation about where ads are working and not working, but it’s often the case that you find previously unknown audiences that a targeting strategy would have missed.
Providers who use a measurement-based testing approach compare actual response data from your placements against each other. Then the firm measures how effective those ads were. You learn more than just what works or doesn’t; after testing, you can distinguish great markets from merely good ones, and good from mediocre. This information is then folded back into the buying process to optimize placements where audience response was best for the buck.
When done properly, a measurement-based testing approach has significant cost savings over an approach based purely on targeting. Because you use small budgets for testing, you only lose money on the test ads that don’t work out. For those tests that do work out, the benefits of that performance can be obtained again and again until a shift in audience response is detected — a shift that may not be detected without this strategy in place.
Audiences move. Testing helps you follow them.
This last point is a significant advantage of a measurement-based testing approach. It’s not easy to tell who’s watching the ads within a particular household or, in the case of streaming, on a particular account. And as the number of channels grows, viewers have more choices and can customize their watch profiles. Target audiences can splinter or drift to other venues. It’s also increasingly difficult to predict what your customers will watch over time. To tell if ads are still landing on their targets, marketers must repeat the research, test conversion, or both. If ad buyers continuously measure response to actual ad placements, they can very quickly shift investment to match the response profiles of the audiences. If you are only targeting, you may miss these shifts in response altogether.
Test small. Scale on what performs well.
The commitment to testing lets you take small, healthy risks. You may pay for some ads that turn out to be low performing. You may also strike gold where you would not have expected success. The truth comes out in the actual measurement of response; a predictive targeting strategy cannot uncover that.
The ability to precisely measure ad performance changes the game and those who are employing testing strategies based on response measurement are gaining significant value while skirting the costs of the old way of doing things.