The definition is the price

You have paid for hours before. The invoice was exact, the pilot shipped on schedule, and the number on the spreadsheet did not move. So when a vendor offers to be paid only if the number moves, it sounds like the fix. It might be. It might also be the same bet with the risk pointed back at you, and from the proposal alone you cannot tell which.

Buyers are asking for it. Futurum's first-half 2026 survey found 27% of buyers favor outcome-based pricing for AI, against 43% who prefer to pay for consumption. Futurum does not publish the sample, so read the split as a direction. Software got there first. Intercom sells its Fin support agent at $0.99 per outcome.

Look at what an outcome is. Intercom counts two kinds of resolution. A confirmed resolution is one where the customer says the answer worked. An assumed resolution is one where the customer leaves the conversation without asking for more help. Both bill at $0.99. That is a defensible line, and Intercom deserves credit for publishing it where anyone can read it. It is also a choice, and it shows where the money in any outcome deal actually sits. Not in the rate. In the definition.

For software, the definition is a paragraph on a help page. For a build engagement it is three decisions nobody enjoys making, and all three have to be made before the work starts: what counts as the lift, what the number would have done without us, and how long anyone watches it. Get one of them wrong and the fee is fiction, in one direction or the other.

The baseline is where these deals quietly fail. A live game's revenue moves with the event calendar and with whatever user acquisition spent last month. Fix the baseline in a quiet month and the vendor collects on the holidays. Fix it in a strong one and the vendor is paid nothing for work that landed. Neither side is cheating. The number was simply never pinned down.

The window fails more slowly. Moloco studied 55 mobile game launches, and one finding stands out for anyone pricing on lifetime value: 87% of eventual top spenders in whale-driven games had not made their first purchase in month one. Grade a monetization change at month three and you are grading it on players who were never going to matter. Grade it at month eighteen and a dozen other changes have landed on top of yours. The honest window is a negotiation between those two errors, and it has to be written down before anyone knows which way the number went.

This is why a firm that offers outcome pricing should turn down a good share of the outcome deals it is offered. Plenty of problems are not legible enough for it. The metric moves for reasons nobody in the room controls, or the feedback loop runs longer than the contract. When we see that, we say so on the first call and suggest a fixed engagement instead. Taking the deal anyway would mean collecting on luck or working for free, and a buyer should not want either.

So if a vendor offers to share the risk, ask to see the definition before the rate. Ask who sets the baseline and from which months. Ask what happens to the fee if the number moves for a reason the work never touched, when the window closes, and who can end it early. A vendor who has thought about outcome pricing will have those answers ready, because the answers are the product.

Outcome pricing does not move risk from you to the vendor. It moves the argument from the end of the project to the beginning, where it is cheaper to have. The rate is the easy part to negotiate. The definition is the price.