Social Rocket
Since the 5th of October 2026, Uber Advertising wants to measure more than just orders placed inside Uber Eats in the US: together with Affinity Solutions, the company offers a measurement solution designed to examine the link between Sponsored Listings and additional sales in physical restaurants. To do this, Uber links its own test and control groups with aggregated, de-identified transaction data from credit and debit cards.
For restaurant chains, this is a bigger change than a new ad format. Until now, a delivery platform could show its contribution to its own sales well – but could only partially answer whether advertising shifts guests from restaurant visits to the app, or actually creates additional demand.
What the reported 40, 27 and 5 percent measure
In an accompanying report, Uber and Affinity cite the following increases during the studied campaigns for people who were exposed to Sponsored Listings: 40% on Uber Eats sales, 27% across all sales channels considered and 5% on in-store restaurant sales. For the six months after the campaign ended, the company reports average increases of 8% in the app and roughly 5% offline.
These percentages must not be added together. The cross-channel figure includes the channels studied and is not a third, additional source of sales. Equally important is the distinction between an observed change in sales and the incremental effect, meaning the business that would not have occurred without the advertising.
The results come from the official announcement by Uber Advertising and the jointly promoted report with Affinity Solutions. A full independent review of the samples, match rates and result ranges is not included in the public release. The figures are therefore company statements, not an effect guaranteed for every restaurant brand.
Why test and control groups matter more than card matching
Card spending can show that part of the studied customer base also later pays at restaurants. But this alone does not prove causation: someone who already likes the brand might have ordered even without the ad. Only comparing suitable groups whose ad exposure systematically differs helps control for this baseline demand.
The IAB and IAB Europe guidelines on incrementality measurement therefore emphasise the importance of clean experiments and name risks from data coverage, selection bias and unsuitable control groups. For Uber, it remains open to public scrutiny how representative the card spending captured via Affinity is, and whether cash, other payment methods or unassigned customers are systematically missing.
What German advertisers can take from this
The launch refers to restaurants in the US; availability of the same solution for German Uber Eats advertisers is therefore not confirmed. The measurement principle is nevertheless relevant for the local market. Even with retail media, audio and streaming advertising, it quickly looks as if an attributed purchase was automatically gained through the ad. How large that difference is, our article on Walmart Connects buyer data in the TV and audio mix explains.
Anyone commissioning a similar study should define in advance the control group design, number of observed locations, sales definition, statistical uncertainty and offline coverage. Our overview of Lowe’s and the role of external incrementality measurement fits here too. And high ad interaction rates are also not the same as additional sales – see the Revlon analysis of Prime Video advertising.
Classification: Uber’s new approach asks the right question: what happens outside the advertising platform? That 5% more offline sales are reported is an interesting finding. Without disclosed uncertainties and independent replication, however, it is not universal proof that delivery app advertising drives restaurant visits.
Image: official company visual from Uber Advertising; symbolic visual, not an illustration of the measurement method.