This Year, Brands Will Seek Out Incrementality

As networks, publishers, and agencies continue to shift to guarantee business outcomes in ad deals (a trend that began earlier in 2019), the concept of “incrementality” will emerge as a key issue for marketers in 2020.

The truth is, we know some customers would have visited a store regardless of seeing a brand’s ad, but we don’t know which audience needs to see an ad to convince them to visit a store. As AI and machine learning technologies advance, savvy brand leaders want to identify the real incremental effect of their campaigns in place of relying on foot traffic and sales data as a masked measure of campaign success.

Advertisers today have an incredibly difficult time distinguishing between those exposed to ads who were already going to visit the store (the natural effect, driven by intent and brand identity) vs. those who visited because of that exposure (the incremental effect, driven by ad sensitivity). Quite understandably, we want to know if our advertising campaigns actually work in changing consumer behavior in our favor.

First, the elephant in the room: selection bias

Incrementality is finally becoming possible to ascertain. Marketers, however, have been reticent to embrace this practice because selection bias — the bias introduced when researchers select a sample for analysis in such a way that the conclusions are driven by the selection process itself — is a worrisome drawback. For example, if brand leaders analyze data on consumers who were already planning to visit their stores, they aren’t going to see the true or real effect of advertising on influencing new visitors and changing consumer behavior. Minimizing this bias is the elephant in the room as well as the biggest problem attribution companies have to address and marketers need to watch out for.

The secret code: causal machine learning

New data-driven tools that minimize selection bias while enabling better decision making are emerging. By tapping into the latest advancements in causal machine learning, these technologies will help marketers understand both how their advertising activations are changing consumer behavior and how to utilize their spend more efficiently as a result. 

Measuring ad impact from granular insights

Powered by these tools, advertising impact will be calculated at a more granular level, and not as an aggregate measure for the entire campaign, as has been the standard. Post-campaign analytics will be hyper-detailed to determine which subgroups in the exposed group were more sensitive (the audience that is generating the most incremental visits) to the campaign message in terms of visitation patterns, demographic breakdowns, and so forth.

This is an exciting development for advertisers as it enables more targeted, more efficient, and more personalized campaigns that plan for the anticipated ad sensitivity levels of recipients.

Getting more bang for your buck

These granular insights improve performance and enable more efficient media dollar allocations. Decoupling organic store visits from incremental store visits will allow marketers to calculate the cost per incremental visit (CPIV) much more easily, so that they can optimize their strategies to lower this metric and save money. For example, they can analyze where best to allocate spend to reach the most sensitive subgroups at the right time and eliminate the areas that provide the least ROI.

Better audience segmentation and targeting

Additionally, these insights, accumulated campaign after campaign, provide an evidence-based framework for audience building and campaign targeting. Marketers will be able to slice and dice performance metrics and segment consumers based on their behaviors and propensity to visit a store to further improve the performance of subsequent campaigns.

Granular, but private

As in 2019, consumer privacy will continue to be a hot-button topic in the new year. So while marketers will crave detail, they will need to choose incrementality tools that are privacy compliant (with laws such as GDPR and CCPA), as well as future-proof should any other legislation come into effect or new industry standards be developed. After all, we expect to see a federal privacy bill on the agenda in 2020.

Data and technology developments will continue to spur exciting new blueprints for advertiser campaign planning. With machine learning making incrementality a more trusted resource, we are inching closer to the holy grail of eliminating campaign waste. I look forward to seeing what the new decade brings.

Antonio Tomarchio is founder and CEO of Cuebiq.