Per-Device vs. Per-User Pricing: Which Model Actually Protects Your Margin
Published by
Throne of Profit EditorialReviewed by
William Hassell
Founder & Chief Editor, Throne of Profit
Most MSP owners pick a pricing model early, usually by copying whatever the shop down the road does, and then never revisit it. Per-device felt simple: count the endpoints, multiply by a rate, send the invoice. It worked when a client meant one person, one desktop, one server closet. But the way people work has changed underneath that model, and the invoice hasn't kept up. The pricing structure you chose for a 2015 client is quietly losing money on a 2026 one — because the thing you're billing for stopped matching the thing you're actually supporting.
Per-device and per-user are the two dominant ways to price recurring MSP revenue, and neither is wrong on its own. The problem is that each one hides a different leak. Device sprawl bleeds you under per-user. Remote work and BYOD bleed you under per-device. If you don't know which leak your model exposes, you're absorbing the cost of support you never priced.
ONE USER, TODAY
[ user ]
├── desktop ┐
├── laptop │ per-DEVICE bills all 4
├── phone (BYOD) │ per-USER bills the person
└── tablet ┘
▲
the gap between what you
BILL and what you SUPPORTOwner symptoms
Your effort per client keeps climbing but the monthly invoice sits flat.
Clients add laptops, phones, and home setups faster than your billing catches them.
Two similar-sized clients have wildly different margins and you can't say why.
Why this happens
The two models bill for different things, and reality has drifted away from both. Per-device counts endpoints, so it assumed a stable, countable fleet. Per-user counts people, so it assumed each person's device load stayed roughly constant. Over the last decade both assumptions broke. A single knowledge worker now runs a desktop, a laptop, a phone, and often a personal device you still end up touching. Remote work multiplied the surfaces you secure and patch without adding a single billable "seat." The model didn't fail on purpose — the ground moved under it, and flat pricing hid the drift until margin started thinning.
Common mistakes
Pricing per-device in a BYOD world, where every personal phone and home laptop you support is unbilled effort you agreed to absorb.
Pricing per-user without a device cap, so one client's four-monitor, three-laptop power users cost the same as another's single-device staff.
Never re-counting the fleet, so sprawl accumulates silently between the onboarding count and today.
Bundling "unlimited devices" into a per-user plan to win the deal, then eating the sprawl for the life of the contract.
Treating servers and network gear as free extras instead of pricing the infrastructure that carries real support weight.
Business consequences
The wrong model doesn't announce itself. It shows up as a slow, invisible margin slide: your team's hours climb with every added device or remote setup while the recurring revenue line holds steady, and the gap comes straight out of profit. A client you onboarded as healthy quietly becomes your worst account, and you only notice when you finally look at hours-per-client. The owner who matches the model to how the client actually works — per-user with a fair device allowance, or per-device with BYOD accounted for — bills for the support they truly deliver, keeps margins predictable, and stops subsidizing their most demanding clients with their leanest ones.
How experienced operators think about it
They stop asking "which model is better" and start asking "what am I actually on the hook to support here." The device is a cost driver; the user is a cost driver; the right model is the one whose billing unit tracks your effort for that client. A firm where every employee runs three devices is a per-user problem waiting to happen without a cap. A warehouse with shared kiosks and few people is a per-device story. Experienced operators also build in a way to re-count reality on a schedule, because any model is only honest on the day you set it — sprawl and remote work make it drift the moment you look away.
Practical actions
Count the real ratio. Pull actual devices-per-user for each client. If it's climbing or wildly uneven, your current model is probably mispriced.
Pick the unit that tracks your effort. Device-heavy, people-light clients lean per-device; people-heavy with predictable loads lean per-user.
Cap or meter the sprawl. If you price per-user, define an included device allowance and a clear rate for anything past it, including BYOD.
Price the infrastructure separately. Servers, firewalls, and network gear carry real support weight — don't fold them in for free.
Re-count on a fixed cadence. Reconcile billed units against actual supported units at least quarterly so drift can't compound.
Questions every owner should ask
If I counted every device I actually support per client, does my invoice reflect it?
Which of my clients would look completely different under the other model?
When was the last time I reconciled billed units against reality, not the onboarding snapshot?
Frequently asked questions
Is per-user always better than per-device now that everyone has multiple devices?
Not automatically. Per-user is often cleaner for typical office clients because it bills the person regardless of how many devices they carry — but only if you cap or meter the device load. Without a cap, a per-user plan is just a per-device plan where you gave away the extra devices for free. And for clients with few people but many shared or specialized endpoints, per-device can still be the more honest match. The right answer depends on that client's device-to-user ratio, not on the trend.
How do I move an existing client from per-device to per-user without a fight?
Anchor the conversation in what you actually support, not in the model name. Show the current device count, explain how remote work and BYOD added surfaces the old invoice never captured, and present per-user as simpler and more predictable for them — usually true. Time it to a renewal or a security upgrade so it lands as part of a larger conversation, not an isolated price hike, and give a clear before-and-after so there are no surprises.
Related articles
Running a Profitable Managed IT Services Business — the pillar.
What Does a Job Actually Cost You? Real Job Costing — the cost side your pricing has to cover.
Am I Charging Enough? How to Know for Sure — the pricing question underneath the model.
Every business has more decisions than time
Whether you need help solving one problem, evaluating a major opportunity, or making a company-changing decision, Throne of Profit gives you consulting capacity on demand.
Purchase only the consulting capacity you need and use it across Weekly Focus, Strategic Focus, Financial Focus, and ThinkTank engagements.