What a 1999 contract still gets right
about outcome pricing.

The whole tech world seems to be in agreement that AI is finally going to make outcome-based pricing a scalable reality.

There's no doubt it's enabling a different conversation. But it should come with a big warning.

AI-native services companies are pushing innovative pricing conversations harder than anyone has before, but they are mostly about consumption, not outcomes. And I'm starting to hear more frequent rumblings about commercials that aren't working in the long term interests of the customer.

All of which sent me back to 1999, when of course I was partying like Prince had advised.

Different solution, and not much AI back in the last century, but the commercial question was exactly the same one being asked now. Who carries the risk, how do you measure it, and what happens if you get the model wrong.

The deal

This wasn't my first outcome rodeo. I'd previously worked in telemarketing, where we were often paid on results. But the 1999 deal was the first outcome-based professional services contract I designed from scratch, with help from Andrew Briggs and Robert Allman.

The client was one of the top financial services companies in South Africa, where we'd been running a managed service for their front office technology.

We'd had two years of building relationships and trust before we even suggested an outcome-based deal. And in truth we suggested it because they weren't entirely sure what we could do for them next.

So we proposed a risk reward contract across their thousand seat contact center operation. If performance improved, we earned more. If it didn't, we earned less.

There was subjectivity in the model, which is what AI will help to remove. But every number traced back to a cell in a spreadsheet, and their financial controller had to approve that cell before we could raise an invoice.

The collar

We agreed one principle up front. Neither party could end up more than 20% better or worse off than if the work had been billed on time and materials. A floor and a ceiling, both ways.

Then we met every month, went through every metric together, and agreed attribution.

Nobody tried to be clever. Nobody went looking for a loophole. Two years of relationship bought us that.

We finished the year between 12 and 15% ahead of what a conventional contract would have paid us. They were delighted, because we did things nobody had written in.

The CFO

The person I think about most now is their CFO.

He didn't just tolerate the deal, he was enthusiastic about it, because he could see the logic and was comfortable with the risk profile.

If a customer ends up paying significantly over fair market value, they'll find a way out at the first opportunity, whatever the paper says.

I checked recently. Twenty-seven years on, that client is still a major customer of NTT DATA.

That's the test I've used ever since. A good deal, outcome-based or not, is built on principles that work for everyone involved.

I've lost some big deals on principle and never regretted one of them.

Martin Dove is the founder of It's the Outcome, and has spent the last decade helping technology and services companies make the shift to outcome-based client engagement.