How the Rules of SaaS Pricing Are Being Rewritten, with Northlane

Seats are no longer the default unit of value

For most of the last decade, SaaS pricing ran on a consistent logic: figure out your packaging, decide which features live in which tier, and then scale price by headcount or users or licenses. The price metric was largely a default. The job of the pricing team was largely to decide how features were bundled and how the customer journey evolved through those tiers.

That logic is breaking down. Not because seats have disappeared, but because AI creates a different kind of value, and that value is increasingly measurable at the output level rather than the input level.

"Because AI can lead you towards such tangible results, we're seeing more of a shift away from seats and more towards outputs and productive work that your AI can do," Saagar said.

What this means practically is that Northlane now puts price metric first in any engagement, before packaging and before price levels. The central question is: what is the unit of value that actually proxies business outcomes for our customers? That question used to have a default answer. Now it requires a real one.

There is a second complication layered on top. In a seats model, discounting a deal was mostly a revenue decision. In an AI world, model costs are real, and your heaviest users can actually be your most expensive to serve. Gross margin is now a live variable in every deal, and pricing strategy has to account for it in ways it simply did not in the SaaS world.

Why most billing tools were not built for this

The theory of consumption-based pricing is straightforward. The implementation is not.

"Being able to track usage from a telemetry point of view and then have that feed to an accurate billing & invoicing model, that's a super, super difficult piece for you to orient your existing subscription system into," Saagar said.

This is exactly what brings companies to Sequence. Billing infrastructure built for fixed recurring charges can technically have usage pricing bolted on, but the result is usually a tangle of manual processes, spreadsheet workarounds, and revenue that falls through the cracks.

Saagar pointed out that this problem was already happening in seat-based models. Companies would provision 50 seats, actually be running 55, and simply not catch it. The tracking was not good enough, and the billing system could not flag it reliably. That was a manageable problem when the gap was a handful of seats. Scale that into a model where every interaction, task, or resolution is a billable unit, and the cost of weak infrastructure becomes significant very quickly.

The other piece companies underestimate is promotional and transitional pricing. Getting early AI adoption often means offering free credits or a free usage tier to onboard customers. But that free tier eventually converts to a paid one, and your billing system needs to handle that transition cleanly, at scale, without manual intervention. Getting customers through that transition is not just an engineering problem, it is a revenue problem.

How to think about usage metrics

One of the frameworks Northlane uses is a breakdown of usage metrics into three distinct categories. Understanding which type of metric you are working with changes the entire pricing conversation.

1. Infrastructure metrics sit at the technical layer: compute, CPU usage, data processing. Pricing here tends to be cost-plus, because the cost driver is direct and visible. This is common at the infrastructure layer of the AI stack. Token-based pricing fits here as well.
2. Input metrics measure effort: interactions, clicks, tasks completed, prompts submitted. These track how much work the product is doing, but they do not distinguish between work that succeeded and work that did not.
3. Output metrics measure productive work done: a support ticket resolved, a document generated and accepted, a workflow completed successfully. This is where Intercom moved with Fin, charging per resolution rather than per seat or per interaction. It is a fundamentally different value proposition to the customer.

The distinction between outputs and outcomes is also worth being precise about. Output pricing says you pay for productive work done. Outcome pricing says you only pay if a defined business result is achieved. That is a much harder model to execute, and the cases where it actually works are narrower than the current discussion suggests.

Outcome-based pricing works in fewer situations than most people think

Outcome-based pricing is attracting significant attention right now. The idea that you only pay when a result is delivered is genuinely compelling as a value alignment story. But Saagar is measured about where it actually holds up in practice.

"The only industry where we've seen this work seamlessly is payments, because there's a closed loop on who gets what and where that percentage of attribution happens."

Chargeflow is a good example of this working well. They recover disputed or lost payments and charge a percentage of what they recover. That model holds because the value chain is short and attribution is unambiguous. The product recovered a specific sum, and the fee is calculated directly from that number.

The moment you move away from that closed-loop scenario, attribution becomes contested. If you are claiming credit for a percentage of a company's revenue growth, the CRO will have a different view of what drove the number. You might believe your product contributed 20% to a million-dollar revenue increase and be entitled to a finder's fee. They will point to the sales team, marketing spend, and renewal effort that happened over the same period.

Saagar's recommendation is to build pricing conversations around ROI and outcomes, but to be careful about directly configuring the invoice around a business case that is two steps removed from a measurable result. The closer you can get to a closed loop, the better. For most AI companies right now, output pricing is where the market is moving, and outcome pricing remains aspirational for all but a narrow set of use cases.

Who owns pricing, and why the CFO's role has grown

At seed and Series A, the CEO typically owns pricing, and that is usually the right call. Pricing cuts across product, sales, finance, and customer success simultaneously, and the trade-offs between those functions require someone with full company context and the authority to make calls across all of them.

As companies scale, RevOps takes on more of the operational ownership, tracking where discounts are happening, what the downstream effects are on compensation, and what the billing infrastructure needs to support. Product marketing contributes the customer value and competitive positioning layer. Finance contributes the margin and forecasting layer.

That finance layer has become notably more important in the AI era, and Saagar is clear about why. Consumption-based models are harder to forecast. Gross margin is no longer stable and separate from go-to-market decisions. The CFO now plays a gatekeeper role on whether a new pricing model gets greenlit or stalls internally.

"They could easily say we're not going for this because I can't predict it, I can't forecast it," Saagar said. "So they really have a decision to make, do they want to help drive towards this world, or are they going to stay and say I don't know what to tell the board?"

His advice to finance leaders is to get closer to the go-to-market reality. There is a top-line growth case for consumption models, not just a margin risk. The companies navigating this transition well are the ones where finance and RevOps are aligned on a shared set of KPIs, even if those KPIs are still being defined. New metrics are emerging alongside ARR and rule of 40 for AI-native businesses, and the finance function has an opportunity to lead that definition rather than wait for it to become standard.

Billing infrastructure has to keep pace with pricing strategy

The thread running through everything Saagar described — credits and drawdown models, output-based pricing, overage capture, margin-aware deal desk workflows — is that your billing layer needs to keep pace with your pricing strategy. None of these work well if the infrastructure underneath is still built for flat-rate subscriptions.

That is the problem Sequence is built to solve. Whether you are running seat-based contracts with usage overages, burn-down credit packages, or a fully consumption-driven pricing model, the goal is the same: your billing should flex with your model rather than constrain it.

If you are working through a pricing transition and want to understand what the infrastructure requirements look like — book a demo.

About Northlane

Northlane is a pricing strategy firm helping VC-backed and fast-growing companies design and implement pricing that scales. They work with executives, product, sales, and finance teams to translate pricing strategy into an executable go-to-market motion. Their clients include n8n, DeepL, and teams backed by leading investors including Atomico, Northzone and Lightspeed. Learn more at northlanepartners.com.