MSP Pricing Models Explained: Per-User, Per-Device, and Tiered Plans (2026)

Two MSPs quote the same 40-person accounting firm. The first quotes $115 per user per month and wins on price. The second quotes $165 and loses. Eighteen months later, the first MSP is bleeding on the account: the firm turned out to run two on-prem servers nobody scoped, the partners demand after-hours support during tax season, and the "all-in" seat never specified what happens when the firm acquires a four-person book of business across town. The cheaper quote was not a better price. It was an unpriced risk transfer. That is the central truth of MSP pricing: the model you choose matters less than how precisely you define what sits inside it.

Per-User Pricing: The Default for a Reason

Per-user pricing charges one flat monthly rate for each employee and covers every device that person touches, typically a laptop, a phone, and email and identity across all of them. It has become the dominant model in MSP pricing for SMB clients because it matches how the buyer thinks. A business owner knows their headcount; they do not know their endpoint count, and they do not want a bill that fluctuates because someone got a second monitor docked laptop. Quoting is fast, invoices are predictable, and the price scales automatically as the client hires.

The model has two failure modes. The first is device sprawl: a "user" who quietly accumulates four endpoints costs the MSP nearly double the management burden of a single-laptop user at the same seat price, so the agreement must cap covered devices per user or define an overage rate. The second is the shared-device environment. A medical practice with 30 staff rotating across 12 shared workstations, or a manufacturer with 80 floor workers and 9 terminals, makes per-user math nonsensical in both directions.

Per-Device Pricing: When the Endpoint Is the Unit of Work

Per-device pricing inverts the model: every managed asset carries its own monthly rate, with workstations, servers, network devices, and mobile endpoints each priced differently. Servers command the highest rates because they carry the highest management and risk burden; workstations sit in the middle; firewalls, switches, and access points price lower. The appeal is honesty. The fee maps directly to the thing that actually generates tickets, patches, and monitoring alerts, which protects the MSP in exactly the environments where per-user breaks: clinics, dental offices, retail, hospitality, and plants where shared machines outnumber people.

The trade is complexity at the quoting stage and friction at every change. Each device added or retired changes the invoice, which means the MSP needs disciplined asset tracking and the client needs tolerance for a bill that moves. Many firms land on a hybrid: per-user for the office staff, per-device for servers, infrastructure, and shared endpoints. The hybrid prices the real environment instead of forcing the environment into a model.

Tiered Plans: Good, Better, Best as a Sales Instrument

Tiered MSP pricing layers two or three bundles on top of either base model: an essentials tier with monitoring, patching, and helpdesk; a standard tier adding EDR, backup, and email security; and a premium tier adding compliance support, vCIO time, and security awareness training. The tiers are as much a sales instrument as a pricing structure. They convert the conversation from "is this expensive" to "which level of protection do we need," and they give the prospect a way to say yes at a lower commitment without the MSP discounting the flagship offer.

The discipline that makes tiers work is refusing to sell the bottom tier to clients whose risk profile demands the middle one. An essentials-tier client who gets ransomware becomes the MSP's reputational problem regardless of what the contract says. The strongest operators now put a security floor under every tier, and many have stopped selling the bare-monitoring tier entirely. Datto's Global State of the MSP research has repeatedly found that competing on price is the top sales challenge MSPs report, and tiering is the structural answer: it moves the comparison from your rate versus their rate to your stack versus their stack.

What the Benchmarks Actually Say

Published numbers in this market are wide bands, not point targets, and they should be read that way. Industry surveys such as Kaseya's Global State of the MSP report consistently place all-in per-user pricing for US SMB clients in the roughly $100 to $200 per month range, with the position inside the band driven by stack depth, SLA aggressiveness, compliance burden, and region. CompTIA channel research places typical total agreement values for a 10 to 50 seat business at $1,500 to $5,000 per month, usually on 2 to 3 year terms, which makes a single signed client worth tens of thousands of dollars over the contract life.

The margin data is the more important benchmark. Service Leadership, which publishes the most widely used MSP financial benchmarks, consistently shows top-quartile firms sustaining adjusted EBITDA far above the median operator, and the gap is driven less by cheaper engineers than by pricing discipline: standardized stacks, enforced minimum seat prices, and agreements that convert nonstandard work into projects. The lesson for a firm setting its own MSP pricing is that the band is wide enough that almost nobody should be anchoring at the bottom of it. A seat priced $30 below market across 300 seats is $108,000 a year of margin given away to avoid a conversation about value.

What Belongs Inside the Seat

An all-in seat price is only as defensible as its contents. The modern baseline includes helpdesk with a stated response SLA, endpoint management and patching, EDR or managed detection and response, email security and MFA management, backup with periodically tested restores, security awareness training, and vendor liaison for the client's line-of-business applications. Cyber insurance requirements have done MSPs a quiet favor here: carriers now demand MFA, EDR, and tested backups as a condition of coverage, which turns stack components that once felt like upsells into items the client's insurer already requires.

Exclusions deserve equal precision. Projects (migrations, office moves, new servers), hardware, software licensing at cost, onboarding of acquired companies, and after-hours work beyond the SLA should be named as billable, in writing, before signature. This is also where qualification tooling earns its keep on the sales side. A prospect who has completed a Which IT Services Does Your Business Need? assessment arrives already sorted into the right service categories, and an IT Needs Assessment Survey captures user count, device mix, compliance requirements, and budget before the first call, which is precisely the data that prevents the mispriced quote in the opening story.

Two pricing mechanics deserve a line in every agreement regardless of model. The first is a monthly minimum: a floor commitment that protects the MSP when a client shrinks below the seat count the onboarding investment assumed. The second is a real onboarding fee. Bringing a new environment under management, documenting it, remediating the worst findings, and deploying the stack is weeks of front-loaded work, and firms that waive the fee to win the deal are financing the client interest-free and hoping the contract outlasts the payback period. Compliance-driven clients, healthcare, finance, defense supply chain, justify their own premium tier, because the evidence collection and audit support never fit inside a standard seat.

Scope Creep: Where Seat Margin Goes to Die

Scope creep in managed services rarely announces itself. It arrives as a new SaaS app the client adopted without telling anyone, a remote hire in another state, a "quick favor" supporting the owner's home network, or an acquisition that added nine users who were never onboarded into the agreement. Each item is individually small; collectively they can consume the margin on an otherwise healthy account. The defense is a counting discipline: a quarterly or semiannual true-up that reconciles billed users and devices against the actual environment, with the agreement specifying that the invoice follows the count automatically.

Creep also hides in vendor sprawl. Clients accumulate overlapping tools, orphaned licenses, and shadow IT, then expect the MSP to support all of it inside the seat. An IT Vendor Consolidation Grader turns that problem into an expansion conversation, letting the client score their own vendor count, overlap, and license utilization and see the waste themselves, which converts unbilled support burden into a billable rationalization project. For prospects still weighing whether outsourced IT is the right structure at all, a Do You Need Managed IT or In-House? decision tool on the website starts the relationship with a fit conversation instead of a rate card, the same upstream-qualification pattern covered in the broader MSP lead generation playbook.

The synthesis is straightforward. Choose per-user for office-centric clients, per-device or hybrid where shared endpoints and servers dominate, and tier the stack so the buyer chooses coverage rather than negotiating rate. Price inside the benchmark band but never at its floor, enumerate the seat contents and the exclusions, and count the environment relentlessly. MSP pricing is not a number; it is a system, and the firms that treat it that way are the ones that show up in the top quartile of every Service Leadership report.