Per-user pricing is still the most common managed services billing model in 2026 — the default for roughly 22% of MSPs, with a national average around $100 to $250 per user per month. But the model is under real pressure, and providers are increasingly pitching outcome-based, value-based, and advisory-retainer structures instead. Buyers evaluating those proposals need to understand what is actually driving the shift, because the honest reason is not the one that appears in the pitch deck.
What's Actually Breaking the Model
Per-user pricing was always a proxy. Nobody believed a user costs a fixed amount to support; the seat count was simply a billing unit that correlated well enough with the labor an MSP expended. That correlation is what automation is severing.
As automation absorbs routine labor — ticket triage, patching, provisioning, password resets, standard troubleshooting — the relationship between seat count and provider effort weakens. Industry analysis expects pure per-device and per-user pricing to increasingly give way to outcome-based and advisory retainer structures for exactly this reason, with skilled engineers redeploying to consulting, architecture, and strategic advisory work as routine tasks get automated.
The commercial consequence is the part worth understanding as a buyer. Under per-user pricing, automation is a margin problem disguised as a virtue. If a provider cuts the labor required to support your 300 seats by 40%, per-user billing means the client eventually captures that saving through competitive repricing. The provider spent capital on automation and hands the benefit to the buyer. Outcome-based pricing decouples the fee from labor consumed, which lets the provider keep the efficiency gain.
That is not a scandal — it is a rational response to a real problem, and a provider that cannot fund automation investment eventually delivers worse service. But it does mean you should read "we've moved to value-based pricing because it aligns us with your business outcomes" as a partial explanation. The complete one includes margin protection. A provider willing to say that plainly is being straight with you.
What the Models Actually Are
Per-user and per-device
Predictable, easy to compare across bids, and easy to audit. The weaknesses are well known: it penalizes you for headcount growth that does not increase complexity, and it gives the provider no financial interest in reducing ticket volume. A provider paid per user has no direct incentive to fix the root cause generating forty tickets a month.
Outcome-based and performance-based
Fees tied to defined results — uptime commitments, security posture measures, a compliance milestone — rather than to inputs. Performance-based variants set fees against SLAs and achieved outcomes.
The theory is sound: align the provider's revenue with what you actually want. The practical difficulty is that it only works when the outcome is measurable, attributable, and within the provider's control. Uptime is a good outcome metric — measurable and largely attributable. "Improved security posture" is a poor one unless you have specified the measurement instrument, because otherwise the provider grades their own homework. And any outcome that depends on your decisions — approving budget for a firewall replacement, for instance — cannot fairly be a provider obligation.
Advisory retainer
A flat fee for strategic capacity — architecture, roadmap, vendor selection, security guidance — usually alongside a separate operational agreement. This is often the most honest framing of what a mature MSP relationship actually delivers once automation has absorbed the routine work: you are buying judgment, not labor hours. The risk is a retainer that quietly becomes a subscription to occasional phone calls, which is why deliverables and cadence need to be written down.
Tiered flat-rate
Fixed monthly fee for a defined scope regardless of seat count. Simple and predictable; the entire negotiation is the scope boundary and what happens when you cross it.
The Clauses That Decide Whether You Benefit
Outcome-based pricing can be better for the buyer or considerably worse. Five contract terms determine which.
1. How is the outcome measured, and who measures it?
Insist on a named instrument and a data source you can independently see. "99.9% uptime as measured by our monitoring platform, with the raw availability report delivered monthly" is a commitment. "Excellent system availability" is decoration. If the provider owns the only measurement and you cannot audit it, you have not bought an outcome.
2. What is the remedy when the outcome isn't met?
This is the single most important question, and the most common failure. A service credit of 5% of one month's fee for a day-long outage is not a remedy — it is a rounding error relative to your actual loss. Look for credits that scale with the severity and duration of the miss, and a defined path to termination for repeated failure. A provider unwilling to attach meaningful consequences to an outcome commitment does not really believe the commitment.
3. What's excluded from scope, and how are excluded items priced?
Outcome-based agreements shift more work into "project" or "out of scope" categories, because the fixed fee has to be protected. Get the exclusion list in writing and get the rate card for excluded work in the same document. The failure pattern: a fixed monthly fee that looks competitive, plus a stream of project invoices that make the effective cost higher than the per-user quote you rejected.
4. Does automation benefit flow to you at all?
Ask directly: as you automate more of our support, does anything about our fee change? A provider may reasonably say no — that is the point of the model — but the answer tells you whether you are buying a fixed price that will look worse over the contract term. A reasonable middle ground is a shorter initial term or a scheduled repricing review, so that efficiency gains eventually reach the market price.
5. What happens at renewal, and what does exit look like?
Outcome-based contracts often run longer than per-user agreements because the provider is amortizing setup investment. Longer terms are acceptable in exchange for price certainty, but not without a termination-for-cause path tied to the outcome misses, documented offboarding obligations, and clarity on data and documentation portability. Our MSP contract negotiation guide covers the broader clause set this fits into.
Comparing Proposals Across Different Models
The practical problem: three bids, three billing models, no common denominator. Normalize them this way.
Convert everything to annual total cost of ownership including the base fee, the projected out-of-scope and project spend based on your last two years of actual work, any onboarding fees, and third-party licensing you will pay separately. Ask each bidder explicitly what they expect to invoice you beyond the base fee in year one; the quality of that answer is itself diagnostic.
Then stress the number two ways. Model 25% headcount growth — per-user pricing scales linearly, flat-rate and outcome-based may not, which is a real advantage if you are growing. And model a bad year: a security incident, a site move, a platform migration. Per-user pricing holds steady while project spend spikes; outcome-based agreements vary enormously depending on how the incident-response scope was written.
Our breakdown of what managed IT services actually cost in 2026 has the underlying per-model ranges to benchmark against.
When Each Model Is Right for the Buyer
Per-user still wins when you need budget predictability tied to a headcount you forecast well, when you want easy year-over-year benchmarking, or when you are small enough that a provider's outcome-based minimum exceeds what seat-based pricing would cost.
Outcome-based wins when you have a specific measurable result that matters more than cost — a compliance milestone with a real deadline, an uptime requirement with quantified revenue consequences — and when you have the internal capability to hold a provider to a measured commitment. That second condition is the one buyers overestimate. An outcome-based contract nobody monitors is a flat-rate contract with extra paperwork.
Advisory retainer plus lean operational wins when you have internal IT capacity and need judgment rather than hands — increasingly the mature mid-market pattern.
The practical takeaway: when a provider proposes moving you off per-user pricing, ask what the remedy is when the outcome is missed and what the rate card is for everything excluded. Those two answers tell you whether the new model transfers risk to the provider or just transfers your savings to them. A provider who has thought seriously about outcome-based pricing will have crisp answers ready; one who is using it as a repricing exercise will not.
Frequently Asked Questions
Is per-user MSP pricing going away?
Not soon. It remains the most common model in 2026, the default for roughly 22% of MSPs, with a typical range of $100–$250 per user per month. Industry expectation is a gradual shift toward outcome-based and advisory retainer structures as automation weakens the link between seat count and provider labor, not a sudden replacement.
Why are MSPs pushing outcome-based pricing?
Partly genuine alignment — tying fees to results the client cares about. Partly margin protection: under per-user billing, automation savings eventually pass to the client through competitive repricing, while outcome-based pricing lets the provider retain the efficiency gain. Both are true, and a provider who acknowledges the second is being straight with you.
What makes an outcome-based contract actually work?
A measurable metric with a named instrument and auditable data source, a remedy proportionate to the loss when the outcome is missed, a written exclusion list with a rate card for excluded work, and internal capacity on your side to actually monitor the commitment.
How do I compare a flat-rate bid against a per-user bid?
Convert both to annual total cost of ownership — base fee plus projected project and out-of-scope spend based on your actual historical work, plus onboarding and separately billed licensing. Then stress-test both against 25% headcount growth and against a bad year with an incident or migration.
← What Managed IT Services Actually Cost in 2026 | Negotiating a Better MSP Contract →