MSP KPIs and Benchmarks

Why KPIs matter before you think you need them

Most technical founders run their MSP on gut feel until something breaks – a big client leaves, payroll gets tight, a tech burns out. By then the numbers that would have warned you were sitting in your PSA the whole time. The research behind Service Leadership's benchmarking is blunt on this point: profitability tracks operational maturity, not size or age. A disciplined two-person shop can out-earn a sloppy twenty-person one on a percentage basis, and the discipline starts with measuring the right things.

The good news is that the MSP industry is unusually well-benchmarked. Between the Service Leadership Index, the Kaseya Global MSP Benchmark, and vendor-published data, you can compare yourself against thousands of peers. Here are the numbers that matter, what "good" looks like as of 2026, and how to track them without hiring an analyst.

Financial KPIs

MRR and recurring revenue percentage. MRR is the single most important number in the business – it is what buyers pay for at exit and what makes your cash flow predictable. Track total MRR, new MRR added, and MRR lost every month. Just as important is the percentage of revenue that is recurring: firms with strong written recurring contracts are worth far more at exit than those living on project work. See pricing models for how packaging drives this number.

Gross margin per service line. Allocate labor and tool COGS to each agreement type and compute margin separately. Typical targets as of 2026:

Never blend product resale into a single margin number – it will mask a managed-services margin problem or make a healthy one look sick.

EBITDA. Per the Service Leadership Index, best-in-class MSPs run roughly 19%+ adjusted EBITDA, the median sits around 9–11%, and the bottom quartile is under 5%. The striking finding: best-in-class firms earn roughly 2.5–3× the median's bottom line at every size. Scale does not fix a broken operating model.

Churn. Top performers keep logo churn under 5% per year. Watch seat-count shrink inside retained clients too – a client who keeps the logo but drops from 40 seats to 25 is churn in slow motion (net revenue retention captures this). Churn is largely set by onboarding quality and QBR discipline, which is why those processes deserve their own attention: see client onboarding and the QBR process.

Service delivery KPIs

These predict the financial numbers six months out. High ticket noise today is margin erosion next quarter.

Tickets per endpoint per month is the diagnostic metric. If you're well above 1, the cause is usually misconfigured environments, weak onboarding, or wrong-fit clients – and the fix is standardization, not more help desk. TruMethods data suggests keeping client stacks aligned to your standards can cut reactive tickets by roughly two-thirds.

Endpoints per tech measures labor efficiency. Below ~250 per tech, your tooling or processes are dragging; above ~400 you risk burnout unless automation is excellent. This number drives your hiring timing.

Utilization is the percentage of paid hours spent on client-facing work. A tech has about 2,080 paid hours a year; at a realistic 65–70% that is roughly 1,350–1,450 productive hours. Resist chasing 80%+ – it leaves no slack for documentation, training, automation, or the P1 spike, and it burns people out.

First response time should be measured against your SLA commitments, with clocks paused on "waiting on customer." A single blended average hides the P1 you missed. See SLA design for setting targets you can actually hit.

Customer feedback should come from surveys tied to ticket or relationship data. Acronis PSA can send customer surveys and report NPS through its service desk. Simplesat is a dedicated option for per-ticket CSAT. Track response rate and score trends by client because a high average based on few responses is weak evidence.

The OML concept: maturity predicts profit

Service Leadership's Operational Maturity Level (OML) framework scores MSPs 1–5 – Beginning, Emerging, Scaling, Optimizing, Innovating – across finance, service, sales, strategy, and compensation practices. The predictive claim, backed by their benchmarking data: profitability tracks maturity level, not headcount. A $5M revenue MSP operating at OML 4 generating around $1M of EBITDA sells at multiples an immature peer of the same size cannot touch.

You do not need to buy anything to use this. ConnectWise publishes the framework for free. Read it once a year and honestly grade yourself; the gap between where you score and OML 3–4 practices is your improvement roadmap. Joining a peer group is the fastest way to get honest external grading.

How a small MSP actually tracks this

You do not need a BI stack. Almost everything above falls out of a PSA you use with discipline:

1. Log all time against tickets and agreements. This is non-negotiable – untracked work makes gross-margin analysis fiction. See ticket management for building the habit.
2. Standardize ticket taxonomy (type/subtype) so ticket counts and trends mean something.
3. Use "waiting" statuses that pause SLA clocks so response and resolution reporting is honest.
4. Sync endpoint counts from the RMM so per-endpoint and per-tech ratios compute themselves. Kaseya's 2025 benchmark found ~95% of MSPs consider RMM+PSA+documentation integration essential.
5. Run a monthly agreement-profitability report – labor hours × loaded cost vs. agreement revenue, per client. This one report finds the client eating your payroll.

MRR, churn, and margin come from your accounting system plus PSA agreement data. If you use Acronis PSA, start with its built-in KPI and profitability reports. Otherwise, a simple spreadsheet updated monthly is enough at first; add a dedicated dashboard product such as BrightGauge only when manual reporting consumes material staff time.

The five numbers to watch monthly when you're small

Under roughly $1M revenue, tracking twenty KPIs is procrastination. Watch these five every month:

1. MRR (total, new, lost) – is the recurring engine growing?
2. Managed-services gross margin – are you above ~50% and trending toward 60%?
3. Tickets per endpoint per month – is your noise floor dropping toward 1?
4. Per-client profitability – which agreement is underwater?
5. Logo churn / at-risk clients – who hasn't had a touch-point in 90 days?

Everything else – utilization, CSAT, response-time attainment – check quarterly until you have a team large enough for the monthly version to be signal instead of noise.

Bottom line

Benchmarks exist so you know whether a number is a problem: 50–60% managed-services margin, sub-5% churn, under 1 ticket per endpoint per month, 250–400 endpoints per tech, 65–70% utilization. But the deeper lesson from the Service Leadership data is that the habit of measuring – time logged, taxonomy standardized, margins computed per agreement – is itself what separates the 19%-EBITDA firms from the 5% ones. Start with the five monthly numbers, run them for two quarters, and let what they reveal set your priorities.