Profit from an auction where most bets lose
Paid search on a fixed-payout model is a game where a lot of your bets lose. The work was building a system that finds the few keyword-and-offer combinations buying clicks cheaper than they pay out — and cuts everything that bleeds. Over about two months it turned ~$111k of spend into ~$139k of revenue: $29k net, 26% blended ROI.
The company
A paid-search operation on affiliate offers, financed by an investor and run for them. The model is simple to say and hard to do: buy search clicks, send them to offers that pay a fixed amount per new paying user, and keep the gap between what a click costs and what the offer pays. The same approach fits any backer who wants paid-search traffic bought against a conversion metric — the industry is interchangeable, the discipline isn't.
The problem
The math only works if two things line up at once, and both fight you.
- The traffic has to be cheap. A payout might be $50, but if the click costs $52 to buy, the play loses money. Profit lives in the spread — pay $35 for what returns $50 — and finding queries that stay that cheap is the whole game.
- The traffic also has to be good. The payout only fires if the user passes the partner's quality bar and actually pays. Cheap clicks that don't convert are just burned budget wearing a discount.
- Most combinations lose. On a live search auction, plenty of keyword-and-offer pairings run negative — the tracker was full of red rows. The blended result only stays positive if the losers get cut fast and the budget concentrates on the few that work.
None of this is a channel problem — the channel was just Google search. What it took was a system: a repeatable way to find winning combinations, prove which ones convert at a profit, and kill the rest before they eat the margin.
What we did
Ads were placed on the keywords competitor brands were bidding on — queries where high-intent demand already existed — and that traffic was routed to the operation's own offers. The aim throughout: find search queries that were cheap to buy and still produced the paid action the offer pays for.
Each play was a combination of keyword, ad, landing page and offer. Dozens ran at once, because you can't know in advance which pairing buys cheap and converts quality. The tracker measured every combination down to clicks, paid actions, cost and ROI.
Queries that bled budget with no result were killed on sight; the ones delivering users who cleared the partner's quality bar got the budget. The blended 26% came from that discipline — not from every play winning, but from concentrating spend on the few that did.
What changed
The channel was never the edge — anyone can open Google Ads. The edge was the system on top of it: intercept demand that already exists, run many keyword-and-offer combinations at once, and hold a positive blended ROI by cutting the losers faster than they eat the winners. That isn't picking the one right bet, and it isn't tied to this industry — it's a portfolio discipline that transfers anywhere paid search meets a conversion metric. That's the layer we built.
Running paid traffic where most campaigns lose money and a few carry the rest?
Write us what hurts. Within 24 hours: an honest answer whether we can help, and the next step if we can.
Get in Touch