Receipt 01 · ROAS 0.7 → 2.4 · 7 weeks · interview with Alina Bondar
Stop paying Tier-1 prices for Tier-3 offers
A sweepstakes advertiser came in running US and UK push at a 0.7 ROAS and asked us to fix the creatives. The creatives were fine. The geography was wrong. Alina ran the rebuild; this is the engagement in her words, numbers from the account's own Voluum reports.
What did the account look like on day one?
Two GEOs, both expensive. US clicks were averaging right around the $0.21 market rate and the payout math never closed: the offer paid $1.80 per lead and needed a conversion every eight clicks to break even. It converted every eleven. That is a 0.7 ROAS engagement signed by someone who believed volume would fix it. Volume does not fix arithmetic.
Why move down-tier instead of fixing conversion rate?
Because sweepstakes follow the GEO, not the polish. The networks' own data all year has pointed the same way: sweeps volume moves through Brazil, the Philippines, India, Bangladesh, and Vietnam, and high-volume Tier-2/3 clicks average under $0.07. The same $1.80-class offer economics that fail at $0.21 a click work at a fifth the price, even with a lower conversion rate.
What did the test actually look like?
Fourteen GEOs across RichAds and Push.House, seeded with sixty creatives from three packs. Every source ID got a written threshold before launch: kill below one conversion per 400 clicks, cap winners at 20% of daily budget. Weeks one and two were deliberately ugly while the data came in. By week four we were running whitelists in eight of the fourteen GEOs and had killed the other six without ceremony.
Where did it land?
Week seven: blended CPC under $0.03 on the surviving whitelists, ROAS at 2.4 on flat spend, and a creative rotation cycling every twelve days. The advertiser's only complaint was that we killed a GEO they liked. It was Bangladesh, it converted, and the payout mismatch on the offer side ate the margin, which the postbacks caught in week three. Sentiment is not a KPI.
What would you tell someone copying this?
Write the kill thresholds before you spend, not after. Our whitelist checklist is the same process this account ran, and the GEO fit checklist covers the tier math with 2026 prices. Everything else was patience.
Running sweeps at Tier-1 prices? The strategy call is free and the GEO math takes about ten minutes of it.
Get a free strategy call