BPS PPC - RC-2 run the two-week hold-out that decides whether halo overturns the -1480 non-branded verdict

What to do

BPS PPC — RC-2 run the two-week hold-out that decides whether halo overturns the -1480 non-branded verdict.

Why

Do: build a two-week geo/product hold-out on the BPS non-branded Sponsored Products lane, per Curry, comparing total sales with and without cold spend. Start it no earlier than 2026-09-18, so it does not overlap the RC-1 bid-cut settling window. Why: halo is the one assumption that could overturn the measured −$1,480.15 non-branded verdict. For halo to rescue it, 15 attributed orders would have to be generating roughly $4,000/month of unattributed sales. That is a large claim with no evidence today, and it is cheap to test. Worth: nobody has sized this yet. No money figure is on file for it. It is 28 day(s) past its ClickUp due date.

If you approve

I carry it out with the usual safety checks and post the proof on the ClickUp task.

If you deny

Nothing is done. The task is closed with your reason and it will not come back in the brief.

Where it is tracked

the ClickUp task

How it ends

Ends when the ClickUp task is closed.

Original wording: BPS PPC - RC-2 run the two-week hold-out that decides whether halo overturns the -1480 non-branded verdict.

Do: build a two-week geo/product hold-out on the BPS non-branded Sponsored Products lane, per Curry [REG-EXT-01], comparing total (not attributed) sales with and without cold spend. Start it no earlier than 2026-09-18, so it does not overlap the RC-1 bid-cut settling window. Why: halo is the one assumption that could overturn the measured −$1,480.15 non-branded verdict (spend $1,881.27 → attributed sales $1,094.15, 171.9% ACoS against a 36.66% break-even, $125.42 per order). For halo to rescue it, 15 attributed orders would have to be generating roughly $4,000/month of unattributed sales. That is a large claim with no evidence today, and it is cheap to test. Worth: nobody has sized this yet — it is an information lever, not a dollar lever. Context needed to execute this without the originating session. Account spine, 2026-08-04 to 2026-09-03: ad spend $2,170.62 · attributed sales $4,615.95 · 57 orders · 47.0% blended ACoS against a 36.66% break-even · net −$478.41. Branded $289.35 → $3,521.80 at 8.2% ACoS (+$1,001.74). The one working non-branded engine is BD ALL SPA-KTA EX at $202.31 → $2,602.50, 7.8%. Break-even ACoS 36.66% is the contribution margin (cm_before_ads 15,171.34 / sales 41,384.50). First-order CAC ceiling $27.93 (AOV $75.47 x CM 37.01%); LTV-adjusted ceiling $29.32.