Many marketers today are implementing attribution systems to help them determine which channels or publishers contribute to conversions. Attribution systems certainly go a long way in helping advertisers understand how best to spend their marketing dollars.
But there’s a bit of a problem, and it’s worth thinking about. Attribution systems can be rather blunt instruments in a specific, but important, respect. Most systems collect all the conversions that result from a campaign and divide up the credit among the publishers that displayed an ad to a consumer who converted. In other words, every publisher receives credit, even if the ads displayed on their sites didn’t actually influence the consumer.
Crediting all publishers, as we have seen in other aspects of the industry, can encourage publishers to optimize campaign execution based on the wrong actions, such as last click, especially if they receive credit for those actions. In such cases, their goal may be to message the consumer as much as possible in order to claim it, which can result in a poor brand experience for the consumer and advertiser.
The question becomes: How can marketers ensure that their commission structures align with their campaign goals? The only true way to ensure that all parties are working toward the same end is to measure the influence the ads have on consumers, and pay only for those that clearly help sway a customer. And that, in turn, begs another question: How can a marketer assess the degree to which an ad has influenced his or her converters?
One way to measure an ad’s influence is to perform an incrementality analysis for the sites on which the ad appears.
Incrementality is measured in numerous fields using A/B tests. For example, it’s used to test the efficacy of new drugs, where a control group receives a placebo, and the other receives the new medication. If both groups respond similarly, the new drug is said to have little to no incremental benefits.
Incrementality measurement is hardly new to advertising. Many marketers assess ad creative and/or placements via A/B tests.
Data scientists who study causality have been exploring ways to simulate the A/B test methodology to measure the true incremental lift, utilizing digital efficiencies. In these scenarios, unviewable ads serve as the control group (finally, a way to put those unviewable ads to good use!). Ensuring that the control group is exposed only to unviewable ads is the foundation of the research. Additional data points available to us today allow scientists to develop an algorithm that ensures sound methodology, eliminating all biases.
The benefits of understanding the true lift on ads on a per-publisher basis are invaluable to marketers. Such insight will enable them to optimize their campaigns to focus on the places of the web that deliver true incremental benefits, and for the first time, align their campaign goals with publisher incentives.