Interviews, insight & analysis on digital media & marketing

Incrementality is the new buzzword in ad measurement, but what does it mean?

By Trevor Testwuide, CEO, Measured

With the threat of recession looming and marketing budget growth stagnating, it is more critical than ever for marketers to have a clear picture of the impact of their media investments. The modern marketing ecosystem is complicated, and every platform is unique. Calculating return on investment (ROI) for a single channel is a challenge. And, for many marketers, measuring across multiple channels and tactics seems impossible.

Traditional methods for media measurement used by ad platforms and attribution vendors were never perfect solutions, but data limitations and increasing restrictions on user-level tracking now plague them with severe accuracy issues. Brands that continue to use these outdated methods are at risk of wasted budget and missed opportunities.

What is incrementality and why is it better?

As marketers search for more reliable ways to track performance and optimise media, old methods are giving way to newer approaches such as incrementality measurement. 

Incrementality in marketing refers to the increase in results over and above what would have happened without the specific marketing activity in question. Take the example of a brand selling a new phone and advertising it on Google. A number of buyers would always have purchased that phone, whether or not they’d seen the advert on Google – maybe they already wanted an upgrade or maybe they saw adverts on other channels that convinced them to buy. The additional phone purchases made specifically because buyers saw the ad on Google represent incrementality, or the incremental contribution that Google ad campaign made to sales.

Incrementality-based attribution addresses the inherent flaws of popular attribution models of the past. Last-touch or “last-click” attribution, used by most digital ad platforms, assigns 100% of the credit for a conversion to the last client touchpoint on the path to that conversion. In the past, this method often resulted in platforms taking more credit than they deserve because it ignores the impact of all the other touchpoints, and it counts conversions that would have happened anyway.

Multi-touch attribution attempts to assign a fraction of the credit for each conversion to multiple touchpoints by building user click paths and applying various types of modelling to determine what percentage of the credit each channel should receive. This approach is complex and difficult to deploy successfully. It also relies on correlation versus causation, meaning it isn’t making a direct connection between media and resulting sales. 

Both of these click-based methods of attribution are also problematic because they rely on tracking and collecting data at the user level. The anti-tracking restrictions introduced by Apple last year, Google’s promise to kill off cookies and the increasing regulations around the world protecting consumer data-privacy, mean these measurement approaches are not just prone to inaccuracy. They are destined to fail. New privacy policies like Apple’s App Tracking Transparency (ATT) have limited the ability of ad platforms to track large segments of their users, causing them to now often under-report conversions.

Ultimately, it doesn’t matter if the platform takes too much credit or not enough. Both result in the advertiser making important media investment decisions based on inaccurate data. Only incrementality connects media investments directly to conversions and, fortunately, measuring it can be accomplished without identifying and tracking users.

So how does incrementality measurement work?

Incrementality is measured through long-proven test-and-control experiment methodology. At the simplest level, the ad or treatment is presented to one segment of the intended audience (the test group) and withheld from another (control group). By subtracting the percentage of people who converted without seeing the ad, from the total conversions made by the exposed audience, we are left with the incremental conversions that specifically resulted from the ad being tested.

Incrementality experiments can vary in complexity, from the simple A/B test described above to multi-factor experiments so elaborate they require the expertise of a trained data scientist. But when properly designed and managed to be scientifically sound, these experiments can utilise data from any number of sources to understand the impact of each channel and tactic across your whole marketing mix.

With every pound of ad spend being scrutinised, connecting media investments directly to business results is vital. Incrementality-based attribution is accurate, holistic, and thanks to automation and the vast amount of intelligence being collected from ongoing experiment results, achievable for any business. There is brilliant data to be taken from any number of ad platforms and sources, but only assessing its impact impartially will provide the transparency needed to make the most effective decisions about media.