How to Improve MSP Margins Without Cutting Service Quality: A Five-Point Audit
To improve MSP margins without cutting service quality, you audit five levers in order: whether your per-user or per-device pricing still matches the work, how much unpriced work has leaked into your contracts, what your tool stack costs per endpoint, how your engineers' time is actually spent, and how much project work goes unbilled. Thinning profit in a growing MSP is rarely one big failure but several small ones compounding, which is why fixing them in sequence matters more than finding a single cure.
Know where your margin sits before you change anything
Margin repair goes wrong when a founder reacts to a bad month rather than to a measured gap, so it is worth placing your business against a benchmark first. SVM A Partners, citing the Service Leadership Index, reports average adjusted EBITDA for MSPs at roughly 11%, while CloudSecureTech notes that the same benchmark records best-in-class providers holding adjusted EBITDA above 19 percent for a sixth consecutive year.
The gap between those two figures is not explained by better salespeople or cheaper engineers, because it usually comes from operational discipline that the average shop has not yet built. At gross margin level, Rallied puts the healthy target at 50 to 60%, with the realised average at 30 to 40%, and if you sit in that lower band the levers below are where the difference is hiding.
Rule of thumb: if your gross margin falls as you add clients, the problem is not volume but the way each new client is priced, scoped and served.
1. Pricing per user or per device: check it still fits the work
Most MSPs chose a pricing model years ago and have not revisited it, which matters because clients change shape faster than contracts do. Per-device pricing tracks the estate you manage but under-recovers when a client's people generate most of the tickets, whereas per-user pricing tracks the support load but under-recovers when one user carries a laptop, a phone, a tablet and a share of several servers.
The practical fix is to take your ten largest clients and calculate, for each, revenue against the tickets, endpoints and engineer hours they consumed over the last two quarters. Where the ratio is clearly out of line, you have a repricing conversation to plan rather than a model to abandon, and the guide on raising prices without losing the clients that matter covers how to hold that conversation without souring the relationship. Many MSPs settle on a hybrid, with a per-user core and per-device charges for servers and specialist kit, so that neither growth pattern quietly erodes the margin.
2. Contract scope leakage: the work nobody priced
Scope leakage is the slow accumulation of work that your engineers do because it is easier than refusing, such as the onboarding of a new starter's personal devices, a line-of-business application that nobody agreed to support, or the after-hours call that was meant to be chargeable. Each instance is small and reasonable, which is precisely why it never reaches the founder until the margin has already gone.
The fix starts with a written list of what is in scope, what is out of scope and what is chargeable, held in the PSA where engineers can see it while working a ticket. Then tag out-of-scope work as it happens rather than reviewing it afterwards, and review the tagged hours monthly by client, so that the pattern becomes visible and the commercial conversation can rest on evidence. You do not need to bill every minute, but you do need to know which clients are receiving free work and decide whether that is a deliberate investment.
3. Tooling sprawl: the cost that grows with every endpoint
Tools tend to be added to solve a problem and almost never removed when a better one arrives, so a stack that started lean can end up with overlapping monitoring, backup, security and documentation products, each billed per endpoint. According to Level.io, citing CompTIA, MSPs now prioritise platforms that reduce administrative overhead rather than add features, with the trend moving towards solutions that replace three to five tools. The same source illustrates the cost: an MSP using 12 tools at $8 per endpoint each can reduce that to $3 to $5 by consolidating to five tools, and a 500 endpoint MSP could save $15,000 to $40,000 per year.
Audit the stack by listing every tool, its cost per endpoint, its owner and the specific job it does, and you will usually find two or three that duplicate each other. Check how your licences are priced as well, since SyncroSecure found that, on two published price lists, the crossover between per-technician and per-device pricing sits near 130 endpoints per technician, which means the cheaper option depends on your ratio rather than on the headline price.
4. Engineer utilisation and endpoint ratios
Utilisation is where service quality and margin are most easily traded against each other, because pushing engineers harder raises output briefly and then lowers it through mistakes, slower tickets and resignations. Outsourceza notes that most MSPs land between 75% and 80% client-facing utilisation as a sustainable target, and the aim is to reach that band through less wasted time rather than longer days.
A more telling measure for an MSP is endpoints per technician. Rallied's analysis of IT to employee ratios puts 350 to 400 endpoints per technician as the standard for a mature, proactive shop, while reactive MSPs run at 150 to 200. If you sit in the lower range, the answer is not more pressure on engineers but more standardisation, automation of repeat fixes and proactive maintenance that stops tickets arriving at all. The article on utilisation targets and traps explains how to measure the figure so that it does not flatter you.
A reactive MSP is not short of engineers but short of standards, and hiring into that gap makes the margin worse.
5. Unbilled project work
Projects such as migrations, office moves and security uplifts are where an MSP should earn its best margin, yet they are often where the most time disappears, because engineers log project hours against the managed service contract or overruns are absorbed to keep the client happy. The managed service then looks less profitable than it is, and the project looks more profitable than it was, so neither number helps you decide anything.
The fix is to give every project its own budget and time code before work starts, to require a change request for anything outside the agreed scope, and to review actual against estimated hours when each project closes. Over a few quarters that review improves your estimating, which is worth more than recovering any single overrun.
Worked example: if engineers log project time against the managed service contract, a client can appear unprofitable and be repriced unfairly while the real leak, an underquoted migration, goes uncorrected and is quoted the same way next time.
The order to work through the levers
Sequence matters, because some fixes produce the data that the others depend on:
- Time capture and project codes first, since without clean data on where hours go, every other decision is a guess.
- Scope leakage second, because it is usually the quickest recovery and needs no client to agree to anything new.
- Tooling third, which is largely within your control and saves money on every endpoint you add afterwards.
- Utilisation and endpoint ratios fourth, which take longer because they depend on standards and automation becoming embedded.
- Pricing last, when you have the evidence to reprice clients fairly and the confidence that the service you are charging for is consistent.
Where Vitori fits
None of these levers is a secret, and a determined founder with a good service manager can work through all five, whether with Vitori or anyone else. The difficulty is usually time, because the founder is still the person who settles scope disputes, approves tool purchases and handles the difficult client, which leaves margin repair as the job that is always next week. Vitori works inside the leadership team, implements the fixes itself and stays accountable until they hold, using its Operational Scale Framework to assess Growth, Delivery and Operations, and the article on what a fractional COO covers in an MSP explains where that role earns its fee. It is not the right answer for every MSP, but if the levers are clear and the capacity to pull them is not, you can get in touch to talk through whether it would help you build a business that runs, and scales, without the founder in every decision.
Common questions about improving MSP margins
What is a good gross margin for an MSP?
A good gross margin for an MSP sits in the 50 to 60% range, according to Rallied's benchmark data, while the realised average across MSPs is nearer 30 to 40%. An MSP below that average should look first at scope leakage, unbilled project time and tool costs per endpoint before assuming that prices are simply too low.
Should an MSP charge per user or per device?
An MSP should choose per-user or per-device pricing by looking at where support effort actually comes from for each client. Per-user pricing suits clients whose people generate most tickets, while per-device pricing suits estates heavy with servers and specialist kit. Many MSPs use a hybrid, charging per user for core support and per device for servers.
Can an MSP improve margins without raising prices?
An MSP can often improve margins without raising prices by recovering out-of-scope work, consolidating overlapping tools, billing project time properly and raising endpoints per technician through standardisation and automation. Repricing is still sometimes necessary, but it lands better once the MSP can show clients exactly what work their contract covers and what falls outside it.
