ITBluPrint

MSP selection without a weighted scoring rubric

Written by Miles Feinberg | Sep 8, 2026, 1:15:53 PM

Three finalists. A long RFP. A polished bake-off of proposals. Then the award meeting starts with "who did people like?" If nobody walked in with a weighted score sheet, you did not run a selection. You ran a popularity contest with better stationery.

MSP selection fails less often from "bad vendors" than from a process that never forced comparable scores. Price gets discussed. A few SLA lines get circled. Someone remembers a strong reference call. Nobody can show how those pieces were weighted against uptime history, transition strength, staffing depth, escalation design, or cultural fit with your operators. The decision still ships, usually on schedule, and the regret shows up after cutover.

What "no rubric" looks like in real committees

You will recognize the pattern if you have sat through it:

  • Each stakeholder scores from a private mental model (IT cares about tools, finance cares about unit price, operations cares about response feel)
  • Late objections appear in the award meeting that were never criteria in the RFP
  • One charismatic presenter reorders the shortlist without new evidence
  • Disqualifying gaps get waived because "we can work that out in onboarding"
  • The written rationale to the board is rebuilt after the vote, not before

That is not a chemistry problem. It is a missing control. Without weights locked before proposals open, every later debate is politics dressed as diligence.

What a usable scoring matrix actually contains

A matrix is not a 40-tab spreadsheet nobody opens. It is a short, shared instrument:

  • Categories that match how MSPs fail you: service desk quality, escalation path, staffing model, transition plan, security and compliance posture, reporting, commercial clarity, cultural fit with how your company escalates work
  • Weights set before proposals are read (example shape only: operations delivery 30, commercial clarity 20, transition 15, security/compliance 15, references and proof 10, culture 10). Your weights should reflect your risk, not a template downloaded the night before
  • 1-5 scores with written anchors so "4" means the same thing to IT and finance
  • Must-pass gates separate from weighted points (insurance, required certifications, hard coverage hours, data residency). Gates kill a bid; weights rank survivors
  • Named scorers and a rule for how disagreements get resolved before the award vote

Price belongs in the model. Price should not be the model.

When to lock the weights

Lock weights after internal discovery and before the RFP hits the street, or at latest before any proposal is opened. If weights change after you see who is winning, you are not evaluating vendors. You are shopping for a score that matches a preferred name.

If leadership changes priorities mid-process (new compliance need, new acquisition, new plant), publish a revised matrix to the committee and, when fairness requires it, allow a controlled resubmission window. Silent reweighting is how protests and internal distrust start.

Selection gates you can run this week

Look for these in your process:

  • A one-page matrix with categories, weights that sum to 100, and pass/fail gates
  • Score anchors written in plain language for each category
  • Independent scoring by at least two functions (for example IT ops and finance) before a consensus meeting
  • A consensus meeting that reconciles scores against evidence, not against who talked longest
  • An award memo that shows final weighted totals and any gate failures waived, with who approved the waiver

References, site visits, and proposal bake-offs still matter. They feed the matrix. They do not replace it.

Where outside help fits

If your team has never run a multi-stakeholder MSP RFP, a vendor-neutral advisor can facilitate discovery, draft the matrix with your real risk weights, and keep the award conversation on evidence. That is the work ITBluPrint does in selection: structure the evaluation so the loudest voice is not the scoring model.

If you want a second set of eyes on whether your next shortlist is actually scored, start with a free advisory session at itbluprint.com/contact-us.