S04 · sector read
The sale itself, read from the retailer’s purchase data.
10 companies · 158 claims captured · 44 about an “Outcome”
Ten retail and commerce media front doors carry 158 claims over 42 captured pages; 44 carry an outcome level, 21 implied and 23 with a stated metric, none guaranteed.
Outcome is the largest claim type here at 40 of 158, ahead of scale 28, capability 27 and new format 25, and the object claimed is the sale itself — sales lift, units sold, revenue per user, ROAS — with five of the ten naming incrementality and three naming new-to-brand orders or household penetration.
Of the 44 outcome claims, 19 name a client, 7 give a number with no named source, 5 point at a method and 13 carry none; 0 cite a third party, against 7 in S03.
The one page that names a design states it without a figure: “We apply advanced matching methodologies grounded in predictive purchase behavior to construct credible test and control groups.” albertsonsmediacollective.com, beside “70 incrementality Studies ran in 2025”.
Figures elsewhere run the other way — “Rowenta: 78% additional sales, 1459% ROAS, +93% CVR” criteo.com is printed with the advertiser named and no window, baseline or attribution rule.
Every figure is read out of the seller's own purchase data, so it travels inside one retailer and not between them; the single page promising independent verification names no provider uber.com.
No page ties a fee to a measured sale: the only published price is a flat annual tier scaled to ad spend skai.io, and the only use of “guaranteed” sits over resonance rather than a result roundel.com.
Of 39 pillars, 23 keep their evidence on the front door, 11 elsewhere and 5 carry none; the Walmart Connect case-study listing and the Instacart Ads subpages returned no body to the fetch, so the read stops at what was served.
Every figure above is recomputed from these files.