The Trade Desk just made a strong performance claim: campaigns using retail data targeting on its platform averaged 41% higher return on ad spend and 25% lower cost per acquisition than campaigns that did not. The figures, published October 5, cover June 2025 to June 2026 and add regional spikes of 125% in APAC and a reported 150% ROAS increase in EMEA.
That sounds like a reason to pour budget into retail data. But the post never reports a sample size, an attribution model, or any matched comparison between the two groups.
What the platform numbers actually show
The Trade Desk pulls these “new platform insights” from its own data. A footnote says the figures cover global platform data from June 2025 to June 2026, excluding Walmart DSP. The company also says over a third of advertisers activated retail data targeting, with adoption up nearly 20% year over year, although the post does not define whether that is relative growth or percentage points.
Spend splits complicate the headline number: 29% upper funnel, 26% mid funnel, and 45% lower funnel. If almost a third of retail data spend sits in awareness-style campaigns, the global ROAS average blends very different campaign objectives.
Why the comparison isn’t proof
Advertisers choose whether to buy retail data. Those that do are more likely to sell products with purchase histories and to run sales-objective campaigns. A brand-awareness campaign with no retail data could end up in the baseline group with low or unmeasured ROAS. Without matching by objective, vertical, budget or channel, the gap describes what happened on the platform, not what retail data caused.
The regional wording adds another issue. The global and APAC lines compare retail data campaigns against campaigns without it. The EMEA sentence says average ROAS “increased by 150%,” which reads as a before-and-after change, not the same two-group comparison. The company does not clarify.
There is also no North America figure in the post’s text, despite the United States being where most of The Trade Desk’s retail data partnerships are concentrated.
An incrementality study, with limited disclosure
The post includes a more serious attempt at evidence: an analysis across 11 brands and 8 retailer datasets using Attain as a retail-agnostic measurement layer. The Trade Desk says every campaign drove statistically significant incremental sales and transaction lift, and 90% of campaigns beat Attain’s new-buyer ROAS benchmarks by an average of 122%.
That is the right question to ask. But the study does not name the brands, retailers, campaign count, study period, holdout design or confidence thresholds. It also averages only the 90% that beat the benchmark, so the 122% figure is not a campaign-wide average.
What to demand before you shift budget
None of this means retail data is weak. It means the buyer has to force better evidence. Use this list in every retail data vendor conversation:
- How many campaigns and advertisers are behind the average?
- What attribution window was used, and who counted the sales?
- Were campaigns matched by objective, vertical, budget or channel?
- Is there a holdout or control group, and what confidence level was hit?
A practical framework: treat platform averages as the weakest evidence, benchmark-beating studies as slightly stronger, and matched incrementality tests or geo experiments as the real decision layer. If a vendor won’t share the denominator, price the claim accordingly.
For The Trade Desk, the timing matters. The post arrives after Q2 revenue grew just 3% to $715 million, shares fell 24% on soft guidance, and about 575 jobs were cut. Retail data is one of the few stories it can tell, but the missing methodology leaves it as a vendor pattern, not a proof point.
Source: PPC Land
