Seat-Based Pricing Is Dying. Outcome-Based Pricing Isn't the Answer.

Most of what I've written this year has been about how buyers find you and how deals get stuck. This one is about something quieter and more structural: the thing you're actually charging for is coming apart.

If you sell software by the seat, or services by the hour, you've built your revenue model on a single assumption — that the value you deliver scales with the number of humans involved. For twenty years that was a safe assumption. In 2026 it isn't.

The bet that's breaking

The market noticed abruptly. In February this year, software valuations took one of the sharpest short-term hits the sector has seen, with hundreds of billions coming off SaaS market caps in a matter of days. The reasoning behind the selloff was simple enough to fit in a sentence: if one AI agent can do the work of five people, charging by the number of people is charging for the wrong thing.

The evidence had been accumulating quietly before that. Atlassian reported its first-ever decline in enterprise seat counts — not slower growth in seats, an actual decline, as customers replaced human licences with agents. Support teams across enterprise accounts have been reporting seat reductions as AI handles more first-line volume. IDC now forecasts that around 70% of software vendors will move away from pure per-seat models by 2028.

Here's the part most commentary misses, and the reason I'm writing this for a mixed audience rather than a purely SaaS one.

The billable hour is per-seat pricing with extra steps. If you run a consultancy, an agency, a law firm, a logistics advisory — your pricing is tied to human time, which means it's tied to human headcount. When AI compresses the hours a piece of work takes, an hourly model turns your own efficiency into a revenue cut. You get better, you earn less. That's a strange incentive to build a business on, and clients are increasingly aware of it.

Software firms are having this conversation loudly. Professional services firms are having it late.

The overcorrection

The industry's answer has been to swing hard toward outcome-based pricing: charge per resolved ticket, per qualified lead, per completed case. Gartner expects roughly 40% of enterprise SaaS to include some outcome-based element this year, up from around 15% two years ago. Intercom's Fin agent, priced at $0.99 per resolution, gets cited in every conference talk as proof the model works.

And it does work — for Intercom. A support resolution is unusually easy to define, count, and attribute. Most B2B value isn't.

Before you redesign your pricing page around outcomes, four problems are worth sitting with:

Attribution

If you charge per qualified lead, who decides what qualifies? If you charge on revenue influenced, you've just invited your customer's analytics team to audit your invoice every month. Outcome pricing works cleanly where the outcome is binary and machine-countable. Everywhere else, it manufactures disputes.

Forecasting

Pure consumption or outcome models make your revenue a function of your customer's activity, which you don't control. Ask anyone who ran a usage-based model through a downturn what that does to a board meeting. You've swapped predictable revenue for someone else's demand curve.

Buyer discomfort

This is the one vendors underestimate. Procurement needs a number to put in a budget line. A pricing model that can't answer "what will this cost us next year?" doesn't feel innovative to a finance director — it feels like risk. As I wrote last month, the buying committee kills more deals than the buyer does, and unpredictable pricing hands them a reason.

Perverse incentives

Charging per resolution rewards volume of problems, not absence of them. Charging per lead rewards lead count, not lead quality. Whatever you price becomes what you optimise for — so choose a metric you'd be happy to be judged on for three years.

What's actually winning: the boring middle

Here's the finding that should shape your decision, and it gets far less airtime than the outcome-based narrative.

When you look at what B2B software companies are actually doing rather than what they're talking about, hybrid models dominate. In a 2026 survey of more than 230 software companies, hybrid was the single most common primary pricing structure — a base platform fee plus a variable component tied to usage, outcomes, or both. Adoption sits somewhere in the low-to-mid forties as a percentage and is expected to keep climbing through the end of this year.

Pure outcome pricing gets the headlines. Hybrid gets the renewals.

The logic is straightforward once you see it. The base fee gives both sides what they need — predictable revenue for you, a budgetable number for procurement. The variable component means your revenue grows when your customer's usage or results grow, rather than when their headcount does. You get expansion revenue that isn't hostage to their hiring plan.

That last point matters more than it sounds. Expansion now accounts for something close to half of new ARR at many B2B companies, and net revenue retention has become the metric investors and boards anchor on. If your expansion path runs through seat growth, and seat growth is flattening across the market, your NRR has a structural problem no amount of customer success effort will fix.

How to actually move

Repricing is one of the highest-risk things a B2B company can do, and most of the damage I've seen came from moving fast rather than moving wrong. A sequence that works:

1. Find the value metric that isn't headcount

What grows when your customer gets more value from you? Documents processed, shipments tracked, policies administered, transactions cleared, endpoints monitored. It should be something they already measure, that rises with their success, and that they'd accept as a fair basis for paying more. If you can't name it, that's the whole project — everything else is downstream.

2. Keep a floor

Resist the urge to go fully variable. A base fee is not a lack of ambition; it's what makes the model financeable for you and approvable for them.

3. Only price outcomes you can measure without argument

If proving the outcome requires a spreadsheet and a negotiation, price the input instead. Usage-based is less fashionable than outcome-based and dramatically easier to invoice.

4. Migrate in cohorts, not all at once

New logos first, then renewals, then a grandfathered path for your existing base. Give long-standing customers a genuine reason to move rather than an ultimatum. Repricing your most loyal accounts first is how you turn a pricing change into a churn event.

5. Treat it as a positioning change, not a finance change

This is where most companies fail, and it's the part that sits in my lane rather than your CFO's. Changing what you charge for changes what you're claiming to be worth. Your website, your sales narrative, your case studies, and your discovery questions all have to shift with it. A pricing model your sales team can't explain in one sentence is a pricing model that will lose deals no matter how elegant the spreadsheet.

The question to sit with

You don't need to have this solved this quarter. But you do need to know the answer to one question, because your buyers are starting to ask it.

If your customer achieved everything they hired you for using half the people they have today, would you make more money or less? If the answer is less, your pricing model is working against your product.

For most B2B companies I talk to, the honest answer right now is "less." That's not a crisis — it's a two-year problem that gets considerably harder if you start on it in year two. The firms handling this well aren't the ones with the cleverest pricing page. They're the ones who worked out early what they're really selling, and had the patience to reprice around it before the market forced their hand.