Offer Testing and Vertical Scouting with Payout Reconciliation

Most offer tests end at forty conversions and a strong opinion. We do not call an arm before it has produced enough conversions for the difference that matters to be visible: for the 20 percent EPC gap that actually changes a buying decision, that is in the region of 250 conversions per arm at the traffic mixes we work with. Where the budget will not reach that number, we say the test is not worth running rather than run it and guess in the readout. The second measurement, which nobody asks for: Whether the network paid for everything it received. Every test we run reconciles our click IDs against what the advertiser settled, monthly, at sub-ID level. A gap the network cannot account for with a named fraud rule is a reason to move volume, and we have moved it. It is also why we favour offers where the advertiser will speak to us directly. A network that will not put us on a call with the advertiser has already told us something useful about the offer. How a scouting sprint works: It starts with a thesis about a vertical and an audience, never with a list of top offers. We take four to six offers that test that thesis, run them on matched traffic with one variable moving at a time, and hold budget per arm fixed so the comparison survives contact with a good week. The readout is a memo: what converted, at what EPC, at what validation rate, what payout we believe you can negotiate given what the test showed, and which offers we would decline on our own account, with reasons. Where the arithmetic is not the arithmetic: An exchange offer paying on registration is priced against a metric that does not pay your bills. We test through to a funded account and to 30-day traded volume, and we will ask the advertiser to restate the payout on that basis. For Indian traffic in particular, an LTV model built on global cohorts overstates the case: gains on virtual digital assets are taxed at a flat 30 percent under Section 115BBH, losses cannot be set off, and 1 percent is withheld at source on transfers under Section 194S. Retail trading frequency responded to that, and so does the second-year value of an Indian funded account. We put both lines into the payback model, and the picture changes, occasionally enough to close the vertical. What you receive: The test plan with stopping rules agreed before launch rather than chosen afterwards. Raw data at click-ID level in whatever format your analyst prefers, not a slide. The reconciliation table. A recommendation with a named next step and the terms to ask for. Where the answer is that none of the tested offers clear, you receive that in the same detail. We have written that memo more than once and it is the one that has saved clients the most money. What is not included: No media budget, which you fund directly. No creative production, no landing pages. No tracker build, which is separate work with its own scope. No ongoing buying against whatever we recommend unless you retain us for it. Who this is not for: Anyone wanting a ranked list of the current best offers in a vertical, which is a thing that exists in forums and is worth what it costs. Advertisers who cannot give us their own funnel data, since we cannot tell you what a conversion is worth when we are only allowed to see the half of it the network reports.