We test the two things that wash out good technicians in an MSP. The first is pace, which here means one queue interrupted by eight or ten client environments instead of one. The second is that the technician is the face of the business on the client’s worst morning. We test both with work, not with questions about the work.
A resume won’t tell you either one, and neither will how well the interview went. A strong candidate can still be wrong for this seat, and that’s a different finding from a weak candidate. So we hand over a small slice of the job and grade what comes back.
One Queue, Eight Client Environments
A technician can solve one hard ticket in a quiet room and stall on a normal Tuesday. The difference is their ability to hold state and switch context without starting over.
Pace at a managed service provider (MSP) has nothing to do with typing faster or staying later. Priorities change mid-task, another ticket ages while you work, and someone else has to be able to pick up the notes cold.
An L2 looks excellent when you hand them a VPN that won’t connect and an hour of quiet. Put the same person on a live queue and the gap appears after the third interruption before lunch. They stop writing things down, and they lose track of which client is which.
That person is a good engineer in the wrong seat, and the two findings have different fixes. Pace matters most in the seats that come open every quarter. Those are L2 and Tier 3 help desk, sysadmins, network and cloud engineers, and the service manager who owns the queue.
The Technician Is the Business on a Bad Morning
In an MSP, the person fixing the problem is also the person the client judges you by. What goes wrong here has nothing to do with charm. It’s an update nobody can act on, or an argument with a client who’s angry and wrong at the same time.
A tech can be right about the cause of an outage and lose the account anyway, because the message reads like blame. The signal we look for is narrow. Can they write an update the office manager can act on, with a next step and a time, and can they correct a client without making them feel managed?
Liking someone in an interview measures none of that. An hour of pleasant conversation measures how well they interview.
Work Samples, Interruptions and Reference Checks
Skip the question about handling pressure. Everyone says yes. All of this comes after sourcing, and where the candidates come from is a separate question with a separate answer.
The Same Three Tests, Every Candidate
The first piece of evidence is a ten-minute work sample. Here is the outage, write the client update. You want a clear next step and an ETA in language the client understands. Grammar isn’t the point.
The second is a walk-through where the interviewer changes the priority partway through the answer. What the candidate drops, and what they write down before dropping it, is the closest you get to watching them work.
The third is a reference check aimed at the behavior instead of the person. “Would you rehire” tells you nothing. “Describe a morning when three of their clients escalated at once” tells you what happened.
Every candidate on the role gets the same sample and the same scenario. A screen that changes shape per candidate produces opinions, and opinions don’t compare. Our screening process runs in a fixed order for the same reason. All of it is candidate evaluation, the C in the FACT Driven Hiring System: Framework, Attraction, Candidate evaluation, Training. We evaluate integrity first, then attitude, then cognitive ability, then personality. Personality comes last because it predicts least.
On a fractional engagement this runs on every open role rather than as a per-role add-on, alongside sourcing, interview coordination and reference checks. The MSPs we work with hire two to five technicians and engineers a year. Across that many seats a test that runs the same way every time compounds, where on a single hire it can look like extra work.
