A customer renews and asks to reduce their licence count from forty to thirty two. Nothing is wrong. They are happy, usage is up, and their support team handles more tickets than last year. They simply need fewer people to do it, because AI is doing part of the work. Under per seat pricing your revenue just fell 20 percent in an account that is more successful than ever.
This is the structural break at the centre of software economics right now, and it arrived faster than most pricing teams expected. Enterprises have reported roughly 10 percent reductions in customer service seats as AI made agents more efficient, and that trend has barely started.
Why per seat pricing is breaking
The model held for two decades because seats were a reasonable proxy for value. More people using the software meant more work getting done, so charging per person tracked the benefit closely enough that nobody argued.
AI severed that link. When software performs work that previously required a person, the number of humans needed drops while the volume of work handled stays flat or rises. Value delivered and seats occupied now move in opposite directions, which means per seat pricing charges least exactly when your product is doing most.
The industry response has been dramatic. A survey of 300 SaaS CEOs in April 2026 found 97 percent planned to retire seat based pricing within two years. The same survey found 94 percent said seat based pricing still aligned with their product's value today, which captures the dilemma precisely: the model works now and everyone can see it will not.
The market has priced this in already. Roughly two trillion dollars of software market capitalisation was erased in early 2026 as autonomous agents undermined per seat licensing assumptions. Investors moved before most pricing teams did.
Why tier jumps feel like a penalty
Tiered SaaS pricing packages value into steps: up to 10 users at one price, up to 25 at another. It is simple to explain and simple to build, which is why almost everyone starts there.
The problem is that customer growth is continuous while your pricing is discrete. A customer moving from 10 to 11 users experiences a 10 percent increase in usage and a 60 percent increase in cost. From their side that is not a price change, it is a penalty for growing, arriving with no warning and no corresponding jump in what they receive.
What makes it worse is the psychology of the moment. They were not shopping. They were doing something routine, adding a colleague, and your billing turned an ordinary Tuesday into a procurement decision. That is the moment competitors get evaluated, and the trigger was your pricing structure rather than anything about your product.
The three responses you get, all bad
When a customer hits a tier wall, there are only really three outcomes and none of them are good for you.
They ask for a discount. You usually give it, because losing them is worse. Now your published SaaS pricing is fiction, your revenue per customer is unpredictable, and word spreads that your prices are negotiable if you push.
They game the boundary. They share logins, deactivate users at month end, or leave people out of the system entirely. Your usage data becomes unreliable, your product gets used badly, and the customer builds a workflow around avoiding your pricing.
They leave. Not always immediately, but the tier jump starts the evaluation, and once someone is comparing alternatives you are in a competition you did not choose to enter.
What to charge for instead of per seat pricing
The work is finding a metric that rises when your customer gets more value. It differs by product and the candidates are usually obvious once you ask the question directly:
- Work completed. Tickets resolved, documents processed, invoices handled, calls answered. This tracks value directly and survives headcount changes entirely.
- Outcomes delivered. Qualified leads, successful placements, completed transactions. Strongest alignment, hardest to measure and attribute cleanly.
- Capacity provisioned. Volume tiers the customer commits to in advance. Predictable for both sides and less precisely aligned with realised value.
- Hybrid seats and usage. A reduced per seat charge for human access plus consumption charges for automated work. Often the most practical transition, since it keeps a familiar structure while capturing agent driven value.
That last option deserves attention because it solves the immediate problem without asking anyone to abandon a model they understand. When a customer removes eight seats but their automated volume triples, the seat reduction is offset by consumption growth, and the conversation stops being a negotiation about shrinkage.
Smooth the curve before you rebuild the model
Moving to a fundamentally different pricing model is a large project. There is a smaller intervention that removes most of the pain and can ship in weeks.
Keep your tiers as packaging, and make the price between them continuous. Instead of jumping from 200 to 320 at user 11, charge the base tier plus a per unit rate for the overage. The customer adding one person pays for one person. When they reach a threshold where the next package is genuinely cheaper, move them and tell them you did.
This preserves everything good about tiers, the simple comparison table and clear feature packaging, while removing the cliff. Most of the anger is not about paying more, it is about paying disproportionately more for a small change. Charge proportionally and the emails stop.
The tier boundary is a support problem too
Something worth checking before you touch anything: how much of your support load already comes from this. Most teams have never counted, and the number tends to be uncomfortable.
Go through the last three months of billing related tickets and tag anything triggered by a tier change. Upgrade confusion, disputed invoices, requests to move users around to stay under a limit, questions about what happens at the boundary. In a lot of businesses this is a meaningful share of all support contact, and every one of those conversations is a cost you are paying to explain a structure you chose.
It also distorts your sales conversations. Salespeople start pre emptively discounting to get customers comfortably inside a tier, or steering them to a larger package they do not need yet so the jump does not happen mid contract. Both behaviours look like good service and both quietly reduce revenue in ways that never appear in a pricing review.
Counting this properly gives you the business case. Smoothing the curve is not only a customer experience improvement, it removes a recurring support cost and stops your sales team discounting around your own structure.
The awkward conversation with your sales team
There is an internal dimension to this that pricing discussions usually skip. Your sales compensation is almost certainly built around per seat pricing, and changing the model without changing the incentives produces predictable resistance.
A rep paid on contract value at signing has no reason to favour a model where revenue accrues as the customer grows. Under usage or hybrid pricing, a smaller initial contract that expands over eighteen months is a better outcome for the business and a worse one for the rep's quarter. They will keep selling the old model wherever they are allowed to, and they are behaving rationally.
The fix is to change compensation alongside pricing, not afterwards. Pay on expected annual value including projected expansion, or introduce a component tied to account growth over the following year. Whatever the mechanism, the sales team needs to earn more when a customer expands, or your carefully designed model will be quietly undermined by the people selling it.
Customer success needs the same treatment. Under per seat pricing their job was retention and upsell conversations. Under consumption pricing, driving adoption directly drives revenue, which is a genuinely better alignment and only works if someone tells them that and measures it.
Changing pricing is riskier than leaving it broken
The market has already repriced software on exactly this question, which we went through in why SaaS valuations fell two trillion dollars. What buyers will pay for changed before most pricing pages did.
Worth stating plainly, because the industry conversation makes migration sound obligatory. Changing a pricing model can trigger churn rather than growth if executed badly, and a flawed model that customers understand is often safer than a better model introduced clumsily.
The failures follow a pattern: too little notice, no grandfathering, a new model that materially increases cost for a visible group of customers, and no clear explanation of what they gain. Any one of those turns a pricing improvement into a churn event, and pricing changes are extremely hard to reverse once trust is damaged.
If you are not ready for a full migration, smoothing the tier boundaries is the high value, low risk move. It fixes the specific moment where customers get angry without asking anyone to accept a new mental model of how your product is sold, and it buys you time to instrument usage properly before the larger decision.
If your billing system cannot support usage based or hybrid SaaS pricing, that constraint is worth addressing before the pricing strategy, since most teams discover the model they want is not something their current billing can express. Our SaaS development and custom software teams build metering and billing systems that handle it. Related reading: subscription versus licensing and SaaS product engineering.
Frequently Asked Questions
How urgent is moving off per seat pricing?
It depends entirely on whether AI is reducing headcount among your customers. In categories like support, sales development, and back office operations the pressure is immediate. In collaborative tools where more users genuinely means more value, seats remain defensible for now. Watch your own renewal data for seat reductions rather than reacting to industry commentary.
What is outcome based pricing and does it work?
It means charging for results delivered, such as a qualified lead or a completed placement, rather than access. Alignment with customer value is excellent and measurement is genuinely difficult, since you must attribute outcomes defensibly and handle disputes about what counts. It works best where the outcome is unambiguous and observable in your own system.
Will usage pricing reduce my revenue?
Not necessarily, and it often increases it over time through automatic expansion as customers grow. The risk is the transition, where accounts currently overpaying relative to usage will pay less. Model the change against real account data before committing, because that group is usually larger than teams expect.
How do I keep revenue predictable with usage pricing?
Use a hybrid model with a committed base covering an allowance, plus overage above it. This gives you forecastable recurring revenue and gives the customer a floor they can budget. Best in class net revenue retention of 120 to 130 percent comes largely from usage and hybrid models expanding automatically as customers grow.
Should I abandon per seat pricing entirely?
Not necessarily, and not quickly. Seats still align with value for genuinely collaborative products where more users means more benefit. The pressure is greatest where AI reduces headcount while delivering the same outcome, since revenue then falls as value rises. Judge it by whether seat count still tracks the value your customers receive.
What is the difference between usage based and consumption pricing?
They are used interchangeably, and the useful distinction is between pure usage, where customers pay only for what they consume, and hybrid, where a committed base covers a usage allowance with overage above it. Hybrid is usually the better commercial choice because it gives you predictable recurring revenue and the customer a forecastable floor.
How do I stop customers gaming usage limits?
Remove the incentive rather than policing the behaviour. Login sharing and month end deactivation happen because the pricing creates a cliff worth avoiding. When cost rises smoothly with usage, the effort of gaming exceeds the saving, and the behaviour disappears without any enforcement.
How much notice should I give for a pricing change?
Ninety days minimum for existing customers, and longer for enterprise accounts with procurement cycles and annual budgets. The notice period matters less than whether anyone is surprised. A customer who learns about a change from an invoice will churn far more readily than one who had months to plan for it.
Talk to our team if your pricing model is losing money on your best customers and nobody wants to be the one to change it.