Charging for performance

When your product recovers revenue that would otherwise have been lost, performance pricing is almost irresistible. You take a percentage of what you recover. The merchant risks nothing. The pitch writes itself and objections mostly disappear.

It's the easiest pricing conversation I've ever had, and it creates a set of problems that I don't think anyone warns you about.

Why it sells so well#

The merchant's risk goes to zero. That's the whole thing. Every other vendor is asking them to pay first and hope, and you're offering to be paid out of money that didn't exist before you showed up. For a merchant spending their own money — which, as I've written, is a completely different buyer than a marketing director — that's an enormous difference.

It also aligns you honestly. You want the same outcome they want, at the same time, in the same direction. There's no version where you're getting paid while they're unhappy.

Where it gets hard#

Attribution becomes the product. The moment your revenue depends on which sales you get credit for, the definition of "recovered" is worth real money, and every merchant will eventually look closely at it. Someone abandons, gets a message, and buys four days later — yours? What if they also got an email? What if they were going to come back anyway? These are genuinely hard questions and you are not a neutral party to them. You are grading your own homework in front of a customer who can uninstall you.

We've been conservative about this on purpose — narrower attribution windows than we could defend, giving away credit in ambiguous cases. It costs revenue. It buys the ability to have the conversation without flinching, and I'd make that trade again.

Your revenue inherits their volatility. Merchant has a bad month, you have a bad month. Merchant runs a huge promotion, you have a great month you didn't cause and can't repeat. You've made your income a derivative of a portfolio of small businesses, which is a more exciting balance sheet than I'd have chosen deliberately.

Costs don't move with revenue. This is the one that actually bites, and it's specific to using humans. Our cost is roughly per conversation. Our revenue is per recovered sale. Those are correlated and not equal, and the gap between them is a conversion rate that varies by merchant, by category, by season, by how good their offer is — none of which we control. There are merchants where we do the work and the economics don't clear, and we found that out one merchant at a time.

It's hard to explain to anyone financially serious. Try describing your revenue quality when it's a percentage of a variable outcome across a long tail of small businesses on a platform you don't own. Every question is a fair question and none of the answers are short.

What I'd tell someone choosing#

Performance pricing is a customer acquisition tool that you pay for out of business model quality. That's not a reason to avoid it — we wouldn't have grown the way we did without it, and for a merchant audience it may genuinely be the only thing that works.

But be clear that you're buying an easier sale with a harder company. The things you give up are predictability, clean unit economics, and the ability to have a simple conversation about what you're worth.

The version I'd want, if I were designing it from scratch, is a floor plus performance. A base that covers the cost of showing up, plus a share of the upside. Merchants accept it more easily than you'd expect, because the base is small and the alignment is still visible. We didn't start there because zero-risk was too good a pitch to give up when we had nothing else.

I think that was the right call at the time and I think it's a decision worth revisiting now that we're not desperate for the pitch.