Why most MSP QBRs drift

A quarterly business review fails in a predictable way: the deck opens with a dashboard, walks through every widget, and ends with the client saying the numbers look fine. Nothing was decided, so nothing changes, and the next quarter looks the same.

The cause is structural. A dashboard answers what happened. A client meeting has to answer what to do next, and those are different questions. If the agenda is organised around the tooling, the meeting inherits the tooling's shape.

The 45-minute shape that works

Four blocks, in this order. The order matters more than the timings: putting evidence before the decision invites the client to argue with the chart instead of the choice.

  • Five minutes — the decisions you are asking for this quarter, stated up front so the client knows what the meeting is for.
  • Fifteen minutes — what changed since last period, with the source line beside each claim.
  • Ten minutes — what the data cannot answer, named out loud rather than smoothed over.
  • Fifteen minutes — the decision itself: owner, date, and what happens if nothing is done.

What to bring as evidence

Two comparable periods, aggregated, from the systems you already run. Not a year of history and not a live dashboard: a client cannot hold twelve months in their head, and a live view invites tangents about this morning's ticket.

Aggregate counts are enough for a decision about direction. Ticket volume, closure, SLA breaches, patch coverage, backup success, critical alerts, repeat incidents and unused licences carry most of the quarterly signal an MSP actually acts on.

What to leave out

Leave out anything you cannot defend if the client pushes. That includes causal claims from correlated metrics, benchmarks against unnamed peers, and financial impact that assumes a rate the client has not agreed to.

Naming the gap is stronger than filling it. Saying that a metric shows the pattern but not the cause reads as competence; presenting a guess as a finding reads as a vendor.

What to send before the meeting

Send the evidence page a day ahead, not the deck. A client who has already seen the numbers arrives with questions instead of surprise, and the meeting starts at the decision rather than at the explanation.

Keep it to one page. If the page needs an appendix, the appendix is the two aggregate source rows, so anyone can check a claim without asking you for the export.

When the client disagrees

Disagreement in a QBR is usually about the meaning of a number, not the number itself. That is why each finding should carry the row it came from: the conversation moves to the interpretation, which is the one you can actually have.

If a client rejects a recommendation, write it down with the date and the reason. Next quarter that record is the strongest evidence you have, and it is the difference between a vendor repeating himself and an advisor tracking a decision.

Turning the agenda into a document

The agenda is only half the work. The client remembers the page you leave behind, and that page has to hold the same discipline: each finding beside its source, each decision with an owner, each gap labelled.

That is the page Sourcerow generates from two rows of the ten-field template. It does not attend the meeting and it does not decide anything — it removes the part where you rebuild the same document by hand every quarter.