Find MSPs

Browse

All States All Cities Best MSPs by State
comparisons

Per-User vs Per-Device MSP Pricing: How to Pick a Model

MyMSPHub Editorial • November 6, 2026

Two honest MSPs can quote the same 40-person company far apart, and neither one is padding the number. One prices per user, the other prices per device, and the gap between the two quotes is arithmetic rather than greed. Understanding per user vs per device MSP pricing is the difference between comparing two providers and comparing two spreadsheets that were never built to sit next to each other.

This guide gives you the one calculation that decides which model fits your environment, a side-by-side comparison you can lift into your own evaluation, a worked example showing the same company priced under both models at two different device ratios, and the definitions you need pinned down in writing before the monthly number means anything. If you want the broader picture on what managed IT costs overall, our managed IT services pricing guide covers the full range of structures. This piece is narrower on purpose: it solves the specific problem of two quotes that will not line up.

Start with the device-to-user ratio test

Before you read another proposal, do this calculation. It takes ten minutes.

Count your users. Count your managed endpoints. Divide endpoints by users. That number is your device-to-user ratio, and it is the pivot on which the entire comparison turns. Then divide the per-user price you are quoted by the per-device price. That second number is your break-even ratio.

  • Your ratio is below the break-even. Per-device pricing costs less. Your people have about one machine each, and per-device billing charges you for exactly that.
  • Your ratio is near the break-even. The models land close, and the real cost difference comes from scope, response-time commitments, and what falls outside the monthly rate rather than from the model itself.
  • Your ratio is above the break-even. Per-user pricing costs less. Once your staff are carrying a laptop plus a desk setup plus a managed phone, per-device billing charges you three times for one person's support load.

With the illustrative rates in the worked example below, $150 per user and $110 per workstation, the break-even is about 1.4 endpoints per user.

The word doing the work in that calculation is "managed." Your ratio is not the number of computers in the building; it is the number of endpoints the MSP will actually support and bill for. The gap between those two numbers is where quotes go wrong. Pull your identity provider's user list, pull the device list out of your existing RMM or antivirus console, pull your asset records, and reconcile all three. Then walk the offices and count what the reports missed: shared workstations, conference-room systems, the machine in the warehouse nobody has logged into since March, anything attached to production or lab equipment.

This is not busywork invented by procurement. Asset inventory sits at position one in the CIS Critical Security Controls, ahead of every other control, because you cannot protect or price what you have not counted. A first proper count can turn up endpoints nobody expected, and every one of them changes your ratio.

What per-user pricing actually buys you

Per-user pricing charges a flat monthly rate per person, covering that person's endpoints up to whatever cap the contract sets, plus the shared services everyone touches: email, identity, security awareness training, helpdesk access.

The logic is that support load follows people, not hardware. One person generates roughly one person's worth of tickets whether they are working from a laptop, a desktop, or both. Password resets, application problems, onboarding, offboarding, and the "my email is doing something strange" category are all human-generated. Under this model, the person who carries a laptop and a tablet costs the same as the person with one desktop, because they call the helpdesk about the same number of times.

Per-user pricing is predictable in a way finance leads appreciate. Your bill moves when your headcount moves, which is a number your business already forecasts. Nobody has to file a change request because someone bought a second monitor and a docking station.

The trap is the word "unlimited." A per-user agreement can be sold as covering unlimited devices per person and still cap covered endpoints in the fine print, and it can treat servers, firewalls, switches, and specialty equipment as separate line items regardless of how many people use them. Unlimited support language without a written scope is a warning sign rather than a benefit, and it belongs on any list of MSP contract red flags. Ask for the per-user device cap in writing. Ask what happens on the fourth device, or the sixth.

What per-device pricing actually buys you

Per-device pricing charges a monthly rate per managed endpoint, which can differ by device class: one rate for a workstation, a higher one for a server, sometimes a separate rate for network gear or mobile devices.

The logic is that cost follows machines. Every endpoint needs an RMM agent, an EDR license, patch management, monitoring, and backup coverage, and those costs are per-machine regardless of who sits in front of it. A server running unattended in a closet generates almost no helpdesk tickets and still needs patching, monitoring, and restore testing every month.

Per-device pricing is transparent in a way technical buyers appreciate. You can see exactly what you are paying for each machine, which makes it straightforward to ask whether that aging workstation is worth the monthly support line or whether it should simply be retired. It can also align better in environments where the machine count is driven by operations rather than headcount: manufacturing floors, healthcare practices with shared clinical workstations, retail point-of-sale, labs.

The trap here is the billable-device definition, and it is the single most consequential term in the quote. Providers differ substantially on whether they count mobile phones, tablets, printers, IP phones, network switches, virtual machines, and powered-off spares. Some count anything with an agent installed. Some count anything with an IP address. Get the billable list by category, and ask what happens to the bill when a device is decommissioned mid-month.

Per-user vs per-device MSP pricing at a glance

Here is the comparison in a form you can lift straight into an evaluation document.

AttributePer-user pricingPer-device pricing
Billing unitNamed person / seatManaged endpoint
Underlying logicSupport load follows peopleTooling and maintenance cost follows machines
Best device-to-user ratioAbove your break-even ratioBelow your break-even ratio
Bill changes whenHeadcount changesAny endpoint is added or retired
Budget predictabilityHigh, tracks a number you already forecastModerate, drifts as hardware accumulates
Cost of extra devices per personZero up to the contract's capFull rate per additional device
Servers and network gearCan be billed separately, so askBilled at a separate device-class rate
Shared or unattended machinesAwkward, no user to attach them toHandled cleanly, each has its own line
Contractors and seasonal staffCostly, each one adds a full seatCheaper if they share existing hardware
Hybrid and remote workforceFavorable, extra home endpoints are absorbedUnfavorable, each home setup adds a line
Failure mode to watch"Unlimited devices" with an undisclosed capBillable-device definition expands quietly
Fits environments likeProfessional services, finance, hybrid officesManufacturing, healthcare, retail, labs

The row to read twice is "shared or unattended machines." A per-user quote has no natural place to put the four shared workstations on your production floor, the two conference-room systems, or the machine that runs your access-control software. They get handled as an exception, and an exception written after signature is negotiated with less bargaining power on your side than one written before it.

A worked example: the same 40-person company at two ratios

Take a 40-person business and price it under both models at two different device ratios. The rates below are illustrative round numbers chosen to show the mechanism, not market quotes. Your actual rates will differ by region, vertical, and scope, which is exactly why you should run your own numbers in the MSP cost calculator rather than borrow anyone's.

Scenario A: a light environment, 1.2 endpoints per user. Forty people, 48 managed workstations, two servers. Mostly one machine per person, a handful of shared and spare units.

  • Per-user at $150 per user: 40 x $150 = $6,000, plus two servers at $250 each = $500. Total $6,500 per month.
  • Per-device at $110 per workstation: 48 x $110 = $5,280, plus two servers at $250 each = $500. Total $5,780 per month.

Per-device comes in about 11 percent lower. The company is not carrying much hardware per head, so paying by the machine costs less than paying by the person.

Scenario B: the same 40 people, 2.4 endpoints per user. A hybrid workforce: 40 laptops, 25 office desktops, 31 managed mobile devices, two servers. Same headcount, same helpdesk load, 96 managed endpoints instead of 48.

  • Per-user at $150 per user: 40 x $150 = $6,000, plus two servers at $250 each = $500. Total $6,500 per month. Unchanged, because the headcount did not change.
  • Per-device at $110 per workstation and $35 per managed mobile device: 65 computers x $110 = $7,150, plus 31 mobile devices x $35 = $1,085, plus two servers at $250 each = $500. Total $8,735 per month.

Per-user now comes in about 26 percent lower. Nothing about the business changed except the hardware attached to each person, and the ranking flipped. That is the entire argument for doing the ratio calculation before you read the quotes rather than after.

Notice what stayed constant in both scenarios: the servers. Servers, firewalls, switches, and other infrastructure can be billed as their own line items under either model, in which case the model debate applies only to the user-facing portion of your estate. If a proposal appears to fold infrastructure into a flat per-user rate, get that coverage in writing, or you may meet the omission at renewal.

The definitions to pin down before the number means anything

Two quotes are only comparable once both sides are using the same words. These are the definitions that move the total, and every one of them should be answered in the agreement rather than on a call.

  • What counts as a billable user. Is it a mailbox, a named individual, a paid identity licence, or an active login in the last 30 days? Shared accounts, service accounts, seasonal staff, contractors, and executives who use email once a week all sit in the gap between those definitions.
  • What counts as a billable device. Get the category list: workstations, laptops, servers, virtual machines, mobile phones, tablets, printers, network switches, firewalls, access points, IP phones. Ask which are billed, at what rate, and which are supported without being billed.
  • The per-user device cap. If the model is per-user, the cap is the term that decides whether the pricing holds. Get the number and get the overage rate.
  • How servers and network gear are treated. Separate rate, separate tier, or included? Physical and virtual priced the same? What happens when you consolidate three physical servers into one host running six VMs?
  • Onboarding and transition fees. One-time setup, documentation, agent deployment, and network discovery can be billed separately, so ask for the onboarding fee as a number. A low monthly rate with a heavy onboarding fee and a low onboarding fee with a higher monthly rate can be the same three-year cost.
  • Adds, removals, and true-ups. How often are counts reconciled? Are additions prorated mid-cycle? Do removals reduce the bill in the same cycle, or only at renewal? Asymmetric terms here, where adds bill immediately but removals wait for renewal, quietly ratchet the invoice upward.
  • What sits outside the monthly rate entirely. Projects, hardware procurement, after-hours work, on-site visits, security incident response, and vendor management are the usual exclusions. Get the list and the hourly rates.

Response-time commitments belong in this list too, because a cheaper model attached to weaker service levels is not cheaper. A per-device quote at a lower monthly rate with next-business-day response is a different product from a per-user quote with a one-hour P1 commitment and 30-minute status callbacks. Our guide to IT support SLAs covers the specific commitments to look for, and you should read the SLA section of both proposals before you compare the totals.

How to normalize two quotes onto the same basis

Once you have the definitions, the comparison itself is mechanical. Work through it in this order.

  1. Fix your counts. One reconciled user number and one reconciled endpoint number, used for every quote. Do not let each provider supply their own count, or you are comparing their assumptions rather than their prices.
  2. Multiply each quote to a monthly total using your counts. Not their sample environment, not a per-seat rate in isolation.
  3. Add the infrastructure lines back in. Servers, firewalls, switches, and anything else billed outside the primary model.
  4. Amortize onboarding across 36 months and add it. Skip it and the cheapest-looking quote may not be the cheapest.
  5. Add a realistic allowance for excluded work. Look at your last 12 months. If you averaged one after-hours incident a quarter and two small projects a year, price those at each provider's stated rates and add them in.
  6. Compare the totals against the service levels, not on their own. Put the P1 response commitment, escalation path, and after-hours posture next to each total before you rank them.
  7. Stress-test at plus and minus 20 percent headcount. Run each total again at 32 people and at 48. The model that survives both directions is the one that fits a business that will not stay exactly this size.

That last step is the one that separates a good decision from a lucky one. Per-user pricing scales cleanly with growth and shrinks cleanly with contraction. Per-device pricing can ratchet: if hardware accumulates faster than it gets decommissioned, the bill drifts upward even when headcount is flat, unless someone is actively retiring endpoints.

So which model should you pick?

Run the ratio, then apply the tiebreakers.

Pick per-user if your ratio is above the break-even, your workforce is hybrid or remote, your headcount is your most predictable number, or your staff carry multiple devices. Then verify the device cap and the server treatment before you sign.

Pick per-device if your ratio is below the break-even, you run shared or unattended machines, your device count is driven by operations rather than headcount, or you want visibility into what each machine costs so you can make retirement decisions. Then verify the billable-device definition and the true-up terms before you sign.

If your ratio lands near the break-even, stop optimizing the model. The two totals will be close enough that the decision belongs elsewhere: response-time commitments, ticket ownership, monitoring scope, security posture, and vertical experience. Providers who cannot answer those in specifics are the wrong choice at either price. Both the NIST Cybersecurity Framework and CISA's best-practice guidance frame security as a set of continuous, evidenced activities rather than a product line, which is a useful test to apply to a proposal: a provider describing what they monitor, how often they test restores, and what triggers escalation is telling you something real, whichever way they bill for it.

One more piece of practical advice: a provider who asks for your endpoint count before quoting is doing the job properly. A provider who quotes a monthly rate off a headcount alone has not looked at your environment, and the number will change once they do. That first conversation tells you more about how they operate than the rate does.

The bottom line

Per-user and per-device MSP pricing are not good and bad options. They are two ways of dividing the same support cost, and the right one is decided by a number you can calculate this afternoon. Count your endpoints, count your people, divide, and compare the result with each quote's break-even ratio.

After that, the model matters less than the boundary around it. What counts as a user, what counts as a device, how servers are handled, what onboarding costs, and what falls outside the monthly rate will move your real cost further than the choice between the two structures ever will. Get those in writing, normalize both quotes to your own counts, and much of the gap between two honest providers can disappear, at which point you are finally choosing on service rather than on arithmetic.

Run your numbers in the MSP cost calculator to get a baseline before you take a single sales call, then compare providers in your state, Texas included, on the MyMSPHub directory of managed IT providers. If you want the full evaluation sequence rather than the pricing slice, the MyMSPHub buyer's guide walks the whole process from shortlist to signature.

Frequently asked questions

How do I compare per-user and per-device MSP quotes?

Normalize both to the same denominator before you look at either headline rate. Count your users and count your managed endpoints, then divide endpoints by users to get your device-to-user ratio. Multiply each quote out to a total monthly number using your real counts, not the provider's assumed counts. Then confirm both quotes cover the same scope: the same servers, the same network gear, the same after-hours coverage, the same onboarding treatment. A per-user quote that excludes servers and a per-device quote that includes them are not comparable until you add the server line back in.

Is per-user or per-device MSP pricing cheaper?

Neither is cheaper as a rule. It depends on your device-to-user ratio and on the two rates you are quoted. Divide the per-user price by the per-device price to get your break-even ratio. Below it, per-device pricing comes out lower because you are paying for a small number of machines. Above it, per-user pricing comes out lower because the extra laptops, tablets, and second workstations stop adding cost. Near the break-even, scope, response-time commitments, and exclusions decide the real cost rather than the model.

What counts as a device in MSP pricing?

That is the definition you need in writing, because it varies from provider to provider. Workstations, laptops, and servers are the obvious ones. Providers differ on mobile phones, tablets, network switches and firewalls, printers, IP phones, virtual machines, and machines that sit powered off in a closet. Some count anything with an RMM agent installed; some count anything with an IP address. Ask for the billable-device list by category and ask what happens when a device is retired mid-month.

Does per-user MSP pricing cover unlimited devices?

Do not assume it does. A per-user agreement can cap covered endpoints per person and treat servers, network hardware, and specialty equipment as separate line items regardless. Unlimited support and unlimited device language without a written scope is a contract red flag rather than a feature. Ask for the per-user device cap, ask what happens on the fourth or sixth device, and get the answer into the agreement.

How do I count endpoints before getting an MSP quote?

Pull three lists and reconcile them: your identity provider's user list, your existing RMM or antivirus console's device list, and your asset or purchase records. Then walk the offices and count anything the reports missed, particularly shared machines, conference-room systems, and equipment attached to production or lab environments. Asset inventory is the first of the CIS Critical Security Controls for a reason, and a first proper count can turn up endpoints nobody expected.

Can an MSP change pricing models mid-contract?

Only if the agreement allows it, which is why the change and true-up terms matter more than the headline rate. Look for how often counts are reconciled, whether adds are prorated, whether removals reduce the bill in the same billing cycle or only at renewal, and what notice is required to restructure the model at renewal. A provider willing to put a mid-term model change in writing with your consent is fine. One that reserves the right to reclassify what counts as a billable device without notice is not.

Which MSP pricing model is better for a hybrid workforce?

Per-user, if your hybrid staff carry several managed devices each. Remote and hybrid staff can accumulate endpoints, a company laptop plus a home docking setup, sometimes a desktop still sitting at the office, plus a managed mobile device. That drives the device-to-user ratio up, which is exactly the condition per-user pricing is built for. The exception is a hybrid business running heavy shared infrastructure such as production terminals or lab equipment, where the device count is driven by machines rather than people and per-device pricing tracks the real support load better.

Free Buyer's Guide

The complete guide to finding and hiring the right MSP for your business.

Download Free Guide

Ready to compare MSPs?

Browse rated providers by state and service type.

Browse the directory