The system
Any MSP can be moved. Most are moved in the wrong order.
Standardise before you measure and you enforce the wrong thing. Automate before you standardise and you automate the mess. The sequence below is the whole product.
Operating principles
Five rules we do not break.
Baseline first, always
We lock the numbers before anything changes. If we cannot measure it today, we do not touch it this quarter — we go and make it measurable instead.
No build without an owner and a metric
Every piece of software we ship has one name against it and one number it is supposed to move. If nobody will own the number, the gap is not real enough to close.
No new roles to keep it alive
If a change only works because someone chases it, it does not work. Coordination roles are the tax that kills MSP margin, and we will not add one.
Two or three things at once, never nine
Concurrent workstreams deepen the trough and nobody can tell which one caused what. We run a short queue and finish things.
We leave
The exit criteria are written in week five: OML has moved, your leads chair the review, and the builds run without us. Dependency is not a business model.
The dip is planned, not survived
We tell you which weeks will be worse, on which team, and by how much. A dip you forecast is a cost. A dip you did not is a crisis.
Four phases
What we do, week by week.
Diagnose
We take a read-only extract from your PSA, RMM, documentation platform and P&L, and we spend time on the service desk watching how work actually arrives. Then we score OML across eight dimensions.
- Twelve baseline metrics locked, with the queries that produced them so you can re-run them yourself
- Ranked gap list — each gap costed in hours, margin or churn risk
- Endpoint and contract reconciliation: what you monitor versus what you bill
- A one-page OML scorecard you can put in front of your board
Sequence
The gap list becomes a queue with an order, and the order is argued out with your leads rather than handed to them. Two or three items are live at any time.
- Build queue with owners, target metrics and a date for each
- Dip forecast — which weeks get worse, for whom, and what we do about it
- Exit criteria written down now, while nobody is under pressure
- What we are explicitly not doing this quarter, and why
Build
We use AI to build the smallest thing that closes each gap, inside the tools your team already opens. Days, not quarters. Then we watch the metric.
- First two or three builds shipped and in daily use
- Each one instrumented, so the metric it owns is visible without asking
- Anything that does not move its number in six weeks gets rebuilt or deleted
- Your team is in the build reviews — they inherit this, so they help shape it
Hold
This is the phase most consultants skip, and it is the only one that decides whether any of it lasted. We stay in the weekly rhythm through the trough.
- Weekly 25-minute number review, chaired by us until your leads take it
- Monthly OML re-score against the locked baseline
- Queue refilled as gaps close — the system keeps running after we go
- Handover when the exit criteria are met, not when the retainer ends
Accountability
The weekly review is the product.
Everything else we build is scaffolding for one habit: a short meeting where every number has a name next to it and that person speaks to it.
It runs 25 minutes. It is not a status meeting, and no one presents. We read the board, the owners explain movement, and we leave with decisions.
Red is allowed. Unowned is not.
A metric going the wrong way is information. A metric nobody answers for is a management failure, and it is the only thing we escalate hard.
One name, not a team
“The service desk” cannot own first-time fix rate. A person can. Shared ownership is the most common way accountability quietly disappears.
The board is visible to everyone
Techs see the same numbers as the leadership team. Nothing corrodes trust faster than a scoreboard only managers get to look at.
We hand over the chair
By the time we leave, your operations lead has run the review for at least six consecutive weeks with us in the room saying nothing.
Scope
What we do not do.
Worth reading before you book. It saves both of us a call.
- Stack replacement
- We will not tell you to rip out your PSA or RMM. Migrations create eighteen months of trough for a benefit that is usually available without them.
- Resold tooling
- We take no vendor commission and resell nothing. If a product is the right answer we will say so and step out of the transaction.
- Staff augmentation
- We are not a NOC, a SOC, or overflow engineers. We change how your operation runs; we do not run it for you.
- Process documentation for its own sake
- We write down what has to be written down and no more. A 200-page runbook nobody opens is bureaucracy with better formatting.
- Open-ended retainers
- Exit criteria are agreed in week five. If we are still here in month eighteen, something has gone wrong.
Start small
Start with the three weeks.
The diagnostic stands on its own. You get the baseline, the gap list and the sequencing plan, and you are free to run it yourself.