Picture a Thursday. Your best T1, two years in, asks for a bump from 48k to 52k. You say no, or you stall, and by Friday they've updated their resume. Three weeks later they're gone for a job paying nearly double, and you're back to hiring. The fix isn't a bigger budget or a better speech. It's a pay band per role, a comp plan your techs can calculate themselves, and a fixed review schedule. Done right, a raise request becomes a data-backed conversation instead of a goodbye notice.
How Wide Should the Band Be? A Rule You Can Apply to Your Market
Set each band's midpoint at your local market median for that role, then give it a 20-30% width. If the local median for T1 is 48k, a 25% band around it lands at roughly 42k to 54k. That puts the 52k request inside the range, so you can grant it without inventing a reason. Those numbers are illustrative. Build your bands from your own market: at least two sources, job boards and a local peer group, refreshed every year. If anyone on the desk sits below the band floor, that triggers an immediate raise to the floor, assuming you can fund it from current cash flow.
The Comp Plan a Tech Can Recalculate in Five Minutes
Add a variable component worth 10-15% of base pay, tied to two metrics the tech can verify every month. First-contact resolution rate and client satisfaction score work because you likely already track them per person. Use a simple table so a T1 at 48k base can see that hitting 90% on both targets earns them an extra 7.2k a year. The key is transparency: they see their own numbers monthly, and they can compute their next paycheck without asking you. If they can't, the plan fails.
| Metric | Target | Payout as % of base |
|---|---|---|
| First-contact resolution rate | 90% | 5-7.5% |
| Client satisfaction score | 90% | 5-7.5% |
| Both | 90% each | 10-15% |
Watch for gaming. Closing tickets without resolving them inflates numbers, and a vague client satisfaction survey invites guesswork. Make the definitions crisp and send each tech their own numbers monthly. In my view, under 5% of base the variable pay stops mattering, so keep it in the 10-15% range.
Schedule Pay Reviews Like an SLA, Quarterly, on the Calendar
Put a 30-minute one-on-one on the calendar for every tech, every quarter, and let the band decide the raise. Tenure and skill growth move someone within the band. An owner's mood on a bad Tuesday does not. In the 52k example, the conversation becomes: the band puts 50k within reach now, and 52k is yours after a certification or two quarters of strong metrics. That is a plan, not a refusal. You need coverage. If the manager is in a review while the desk is uncovered, tickets back up and the whole exercise costs more than it saves. Book a backup person for those blocks.
When the Band Won't Save You: The Limits You Need to See
This system breaks under specific conditions, and you should check them before committing: - Market data is stale or pulled from one source. Refresh annually, from at least two. - Variable metrics get gamed or client satisfaction is vague. Tighten definitions or the plan rewards the wrong behavior. - You won't fund floor raises. If current pay sits far below the band, the first raises may be unaffordable. The band exposes the gap; it doesn't pay for it. - The band is below market because your MSP is small and can't match poaching salaries. A band below what a junior sysadmin role pays elsewhere won't stop someone leaving for double. At very small shops, this is the hard truth.
Monday: Pull Your Local Salary Data and Draft the T1 Band
Spend two hours collecting T1 salary data from at least two local sources. Calculate the midpoint, build a 20-30% band around it, and check where your current techs sit. If anyone is below the floor, that's your first raise. Then write the comp table and put the quarterly reviews on the calendar before the quarter starts. You now have the tool to turn pay requests into a predictable, fair system. The tech who asks for 52k gets an answer backed by numbers and a path to get there. That's how you keep someone from becoming a farewell.
mspoutpost.com/articles/the-52k-request-that-ended-in-a-farewell-and-the-band-that-would-have-kept-it
Written by Megan Friesen
MSP Outpost publishes information, not legal, financial, insurance, tax or compliance advice, and reading it creates no professional relationship. For decisions about your business, talk to a qualified professional.
