How incrementality solves for advertising’s attribution problem
Advertising is often described as part art, part science. (Although the creatives and the data wonks are sure to tell you it’s more one than the other.)
With the advent of Big Data, however, advertising has been able to combine the artistic and scientific parts of the practice and develop a better sense of what kinds of messages resonate best with different kinds of audiences.
Contemporary advertising campaigns span so many different mediums, though, that it’s a challenge to determine the effectiveness of each individual component. Thus, the industry has developed incrementality, a practice for measuring exactly that.
So, what is incrementality?
One of the biggest issues in digital advertising is attribution. That is, when evaluating the effectiveness of an advertising campaign, how much credit should be attributed to the various aspects of the campaign?
A contemporary ad campaign includes a variety of messages disseminated across a host of mediums, including everything from TV to email to Snapchat. Given how wide-ranging campaigns have become, it can be difficult to determine how each discrete channel contributed to the campaign’s overall success.
Incrementality aims to solve for that and measure how effective each distinct part of a campaign is.
How does it work?
Using the good old-fashioned scientific method. Much like you were taught in your middle school science class, conducting a sound incrementality assessment requires creating a control group and an experimental (or test) group — although in the context of advertising, this is often referred to as A/B testing.
In an incrementality study, a group of consumers is randomly sorted into two groups: groups A and B. Group A, the test group, will be exposed to different elements of the ad campaign. Group B, the control group, will not. The advertiser then studies how members of each group behaved — specifically, whether they interacted with the brand being advertised.
Did the people exposed to the ad campaign visit the brand’s website at a higher rate than the control group? If so, what was their reaction at each stage of the ad campaign? Did they buy the product being advertised at a greater rate?
Answering those questions helps determine the incremental value of the campaign.
So why does it matter?
For advertisers, incrementality matters because it helps them better allocate their advertising budgets. And for publishers, incrementality matters because it helps prove that their ad inventory provides value to brands.
Using incrementality, brands can identify which tactics of their campaign are most effective (and which are lagging behind) and thus be more efficient about how and where they spend their ad dollars. Publishers use incrementality to show brands that their ads have a demonstrably positive effect for brands, and thus charge higher rates.
How does it compare to other forms of measurement?
Without incrementality, advertisers have to resort to less precise forms of measurement, such as last-click attribution. As the name indicates, last-click attributes the success of the campaign to the very last action a consumer makes before converting.
Say, for instance, a brand runs a multi-faceted campaign that includes CTV ads, email blasts to existing customers and display ads across a host of different websites. A consumer, John Doe, watches one of the CTV ads and learns about the brand. A day later, he sees a display ad for the same website, clicks on it, gets sent to the brand’s website and learns even more about the brand. While on the website, he gives the brand his email address. A week later, he gets an email from the brand advertising a discount, and he clicks on it and finally decides to buy the product being advertised.
Using last-click attribution, the email ad would receive all the credit for persuading the consumer to buy the product, when in fact it was all aspects of the campaign, working in conjunction, that collectively convinced the consumer to buy.
Incrementality is a more involved process, but it solves advertising's attribution challenges by taking everything into account and measuring its unique value.