HR for Managed service providers: what to set up first

An MSP with twenty-eight staff and forty client contracts begins to feel the strain. The founder still knows every firewall. Two of the best engineers are covering three clients each. A new hire from a product company asks, in week two, what exactly L2 means here, and nobody answers the same way twice. This is the moment to put structure around people before the next big contract arrives.

Here is a build order for founders running managed service providers. More about sister groups, from support desks to data centres, is on the IT infrastructure and managed services hub.

Write what an L1, L2 and L3 engineer owns in your firm. Do it by ticket type, tools and decision rights, not by years of experience.

Tier name and role title (service desk analyst, systems engineer, senior engineer, solutions architect):
Ticket types owned alone (password and MFA resets, printer queues, patch deployment, backup job failure, firewall rule change):
Tickets that must be escalated, and to whom, within how many minutes:
Tools with write access (RMM console, PSA tool, hypervisor, firewall, directory):
Changes this tier may approve without the change advisory call:
Skills shown to move up (a runbook written, a client migration led, a post-incident review owned):

This single page is the base for hiring, pay bands, training and appraisals. See role design and job descriptions for MSPs for how to turn it into role cards.

The first two weeks with a new client decide whether the engineers like the account. Most MSPs improvise. Instead, write a checklist: network diagram, admin credentials in the vault, backup status, patch baseline, licence counts, renewal dates, the five most fragile systems and the client's difficult people.

• Pair a primary and a secondary engineer to each client from day one.
• Hold a one-hour knowledge transfer with the previous provider or the client's own admin.
• Record the odd things: a legacy ERP on a Windows 2012 server, a branch office with one link, a director who calls the CEO on every outage.
• Set the first quarterly business review date before the first month ends.

Billable utilisation tells you how many hours were charged. Alone it tempts people to close tickets fast and ignore root causes. Pair it with reopen rate, first-contact fix rate and client satisfaction notes from the quarterly reviews. Share all four with the team.

An on-call rota written by the founder from memory feels fair to him and not to the team. Put names, hours and pager counts on a shared sheet. Add a skill matrix: rows are engineers, columns are tools and platforms (Microsoft 365, Azure, VMware, Fortinet, backup suites), cells show none, learning, independent or can teach. The matrix shows where one person is a single point of failure.

1. 1

   Review the pager monthly

   Count pages per engineer. Move difficult clients around.

2. 2

   Rest after a bad night

   State what a night call earns in a later start or a day off.

3. 3

   Cross-train the weak cells

   Pick one tool per quarter and pair a learner with an owner.

4. 4

   Publish the ladder

   Show what moves a person up the matrix and into the next tier.

For the matrix design, read competency and talent management. For how to measure people after, see performance management for MSPs.

Screens, chairs, cold server rooms and long shifts are small things that wear people down. Use the workplace health, safety and wellbeing page for ideas such as break rules for night desks and eye checks.

MSP hiring fails when it copies product-company habits. A whiteboard puzzle says little about how a candidate handles a furious office manager whose printer queue froze at month end. Run a ticket-queue exercise instead. Give five mock tickets, mixed in severity, and watch how the candidate triages, what she asks, what she escalates and how she writes the client update. Add a scenario on a failed overnight backup and a phishing click.

Look for steadiness, tidy notes and curiosity about root cause. Keep a pay band for each tier ready before the offer, so the conversation stays about role and growth, not about haggling.

Do not start with an engagement survey, a competency framework with forty skills or a bonus plan with six metrics. Start with tiers, runbook, rota and matrix. Add the rest when those four are used weekly.