Operational maturity level
You cannot argue with a baseline.
OML is one number between 1 and 5, scored across eight dimensions of how your MSP runs. It exists so that six months from now the conversation is about a delta, not a feeling.
The five levels
What each level actually means on a Tuesday.
Maturity language is usually abstract enough to be useless. Here is what each level looks like on the floor.
Most MSPs we baseline land between 1.6 and 2.4. Level 3 is where margin starts to compound. Level 5 is a direction, not a destination.
Eight dimensions
Where the number comes from.
Each dimension is scored 1–5 from your own data, not from a questionnaire about how you think you are doing. The overall OML is the weighted mean.
Service desk flow
How work arrives, gets triaged and gets closed. Ticket volume per endpoint, first-time fix rate, reopen rate, time in queue before first human touch.
Standardisation
Share of endpoints on a known-good configuration, number of live stack variants across clients, and how many exceptions exist because someone once asked nicely.
Automation and centralisation
What percentage of reactive work is already handled without a human, and how much of your alerting still pages a person for something it could close itself.
Data integrity
Whether the PSA, the RMM and the documentation agree with each other. Agents versus billed contracts, asset records versus reality, time entries versus the day.
Project delivery
Realisation rate, estimate accuracy, on-time completion, and how much project work is quietly being absorbed by the managed services contract.
Client management
QBR cadence actually held versus scheduled, contract margin spread across your client base, and how many clients you are losing money on without knowing it.
People and capacity
Technician utilisation, escalation depth, time for a new hire to close their first ticket unaided, and how concentrated your knowledge is in one or two heads.
Financial discipline
Service gross margin, effective rate per endpoint, cost to serve per seat, and whether anyone can answer those questions without opening a spreadsheet.
The baseline
Twelve metrics, locked before we change anything.
We hand over the queries as well as the answers. If you cannot re-run the baseline without us, it is not your baseline.
| Metric | Why it is on the list | Source |
|---|---|---|
| Service gross margin | The single number that tells you whether the operation works. Everything else is an explanation of this one. | Finance |
| Reactive hours per endpoint | The clearest read on whether standardisation and automation are real or aspirational. | PSA |
| Tickets per endpoint per month | Rising volume on flat headcount is the trough arriving whether you planned it or not. | PSA |
| First-time fix rate | Measures triage quality and knowledge depth at the same time. Hard to game. | PSA |
| Mean time to resolve, by priority | Aggregate MTTR hides everything. Split by priority it shows where the queue actually jams. | PSA |
| Endpoint standardisation | Share of devices on a known-good config. The ceiling on every automation you will ever build. | RMM |
| Agent-to-contract variance | What you monitor versus what you bill. Almost always found money, in one direction or the other. | RMM + PSA |
| Technician utilisation | Read alongside margin, not alone. High utilisation on unbilled work is a worse problem than idle time. | PSA |
| Project realisation rate | Where MSP margin most often leaks, because project overrun is absorbed rather than reported. | PSA + Finance |
| Contract margin by client | Ranked. The bottom three usually explain more than the top twenty. | Finance |
| Onboarding time to first solo close | How transferable your operation is. Long ramp means the process lives in people, not systems. | PSA + HR |
| Documentation accuracy | Sampled, not self-reported. We pull records at random and check them against the device. | Sampled |
Read-only access throughout the diagnostic. We do not need write permissions to measure you.
The scorecard
One page, in front of your board.
The diagnostic ends with a single sheet: your OML today, the score on each dimension, the gap list ranked by what it costs, and the sequence we would run.
It is deliberately short. A 60-page report is where accountability goes to die.
Straight answers
The questions we always get.
- Is this a standard framework
- No. Operational maturity models are common in the MSP channel and we are not claiming to be any of them. This is our own scoring method, and we publish the dimensions and metrics above so you can judge it before you buy it.
- How long until the number moves
- A tenth of a level is realistic inside a quarter on one or two dimensions. A full level typically takes nine to fifteen months, and it gets worse before it gets better.
- What if my data is a mess
- It usually is, and that is itself a finding — data integrity is one of the eight dimensions. We score what is there and tell you what we could not measure and why.
- Do you need admin access
- Read-only, scoped to the platforms in the baseline, for the duration of the diagnostic. We will work from exports if your policy prefers that.
- Can we run the diagnostic and stop there
- Yes, and some clients should. You keep the scorecard, the gap list, the sequencing plan and the queries. There is no clause that makes the rest of it compulsory.
Start small
Get your baseline in three weeks.
Fixed fee, read-only access, one page at the end. Then you decide whether closing the gaps is worth the dip.