Anyone else seeing CTV sellers push performance guarantees tied to MMM and in-platform lift? In the last 10 days, three partners pitched us CPM floors with retail media audience overlays, and I’m debating whether to spin up a dedicated CTV pod or keep it inside our growth team. If you’ve shifted 10–15% from social into CTV lately, how are you staffing and packaging it for clients?
And > If you’ve shifted 10–15% from social into CTV lately, how are you staffing and packaging it We kept it in growth but spun a 2-person CTV strike team for 90 days and only took guarantees tied to retailer clean-room sales with geo holdouts — not in-platform lift — with a single deck/IO and a clawback if cost-per-incremental-order missed by 15%; it worked, but only after we insisted on weekly log-level files and 1.5x freq caps. Are your sellers agreeing to geo holdouts or just MMM reads?
Keep it under growth, but carve a short-term CTV pod that owns one “guarantee rubric”: a standardized MMM input sheet, a pre-registered lift design, and log-level delivery or no deal. Push sellers to pay misses in cash credits tied to an incrementality KPI you define; otherwise the “guarantee” is just a shiny sticker. @tdavis did your team get episode-level logs with retail matchback SLAs, or only rolled-up reports?
We’ve started forcing any CTV seller offering “CPM floors” with retail media overlays to back-test their guarantee against our last MMM run and sign an IO clause that converts a miss into invoice credits, not bonus impressions. That let us shift about 12% from social without spinning a new pod, though if MMM cadence is slow, we trigger on a 7-day non-brand search lift instead.
, seeing the same “CPM floors” + retail overlays in the last 10 days — before spinning a pod, I’d run a 60-day pilot owned by one PM who also covers paid social, with a single guarantee scorecard that locks MMM inputs, defines lift windows, and sets a minimum audience match rate (≥70%) or it doesn’t qualify. If they won’t share methodology or match, treat it as brand, not performance. Are your overlays clearing 70–80% match?