MSP Selection Blog | ITBluPrint

Ask MSP Finalists for a Redacted SLA Report | ITBluPrint

Written by Miles Feinberg | Oct 27, 2026, 1:24:27 PM

Every MSP finalist can show you an SLA page: a 15 minute response on critical tickets, resolution targets by priority, a compliance number up top. That page is a promise. What you need to see is how the vendor reports against it in a month that didn't go well.

So before final scoring, ask each finalist for one thing: a real monthly SLA report they sent a current client, with names and identifying details redacted. It should be the report a client actually opened last month, not a mockup built for your RFP.

Our post on beating "industry standard" SLAs covers what the numbers should be. This one is about the evidence: whether a finalist's reporting would let you catch a miss before your users do.

Why the sample tells you more than the SLA page

Sales writes the SLA page. The monthly report comes out of the service desk and its reporting tool, so it shows what the vendor actually measures and what they're comfortable showing a client. If the sample is thin, the reports you get after signing will be thin too. If a finalist can't produce one at all, you've learned something while you still have options.

The checklist

  1. Response and resolution, split by priority. Separate numbers for P1 through P4 (priority 1 is the most critical, priority 4 the least). A single blended average lets a pile of quick password resets hide a slow outage. More on that in how averages hide real resolution times.
  2. Breaches listed one by one. Each missed ticket should show its priority, how far past target it ran, and why. A compliance percentage alone doesn't tell you which tickets missed.
  3. Trends across several months. Look for at least three months side by side, or a trend line, so you can tell whether misses are rare or creeping up. Anyone can pick one good month.
  4. Clear clock rules. The report should say when the SLA clock pauses (waiting on the client, for example) and how paused time shows up.
  5. Reopened tickets counted. A ticket that's closed and then reopened should count against the original, not start a fresh clock.
  6. Tickets AI handled, flagged separately. If the MSP uses AI for triage, password resets, or first replies, the report should show which tickets AI resolved on its own, which it handed to a technician, and how each group performed. Instant AI acknowledgments can make response times look great while the hard tickets wait in the same queue. Check whether tickets AI closed get reopened more often, too.
  7. The same format every month. Ask whether this is the standard report every client gets. A format that changes when a number looks bad is a warning.

How to ask for it

Send the request to every finalist in identical wording so you're comparing like for like. Ask for a report from a client close to your size and service scope, sent within the last few months, and say up front that redaction is expected. Then score what comes back in your evaluation matrix next to references and pricing.

When you walk through the sample with each finalist, ask them to explain one breach in detail. The way a vendor talks about a miss tells you as much as the numbers do.

Talk it through

If you'd like help writing the request and scoring the reports that come back, book a free advisory working session with ITBluPrint. We run vendor-neutral MSP selection and don't sell managed services.

Book a free advisory working session