Managed services providers are one of those categories where the gap between the best and the worst is enormous. A good MSP becomes a genuine extension of your IT team — responsive, proactive, and invested in your outcomes. A bad one bleeds you with scope creep, hides behind SLAs designed to protect themselves, and delivers the bare minimum while charging for the premium tier.

I’ve seen both ends of the spectrum, and I’ve helped organisations transition between MSPs more times than I’d like. Here’s what I’ve learned about evaluating them without getting burned.

Start With Your Actual Requirements

This sounds obvious, but most companies skip it. They go to market with a vague brief — “we need someone to manage our infrastructure” — and then evaluate proposals based primarily on price. That’s how you end up with a provider who technically meets the contract terms but doesn’t actually solve your problems.

Before you talk to anyone, document:

What you need managed. Be specific. Servers, networking, endpoints, cloud infrastructure, security monitoring, backup and recovery, applications. If it’s not in the scope, it won’t be in the service.

Your current pain points. What’s going wrong today? Slow incident response? Inconsistent patching? No after-hours support? These should be explicitly addressed in any proposal.

Your expectations for response and resolution. Not SLA numbers — actual expectations. When something breaks at 2am, what do you expect to happen? Do you need someone answering a phone, or is a ticket acknowledgement sufficient?

Your growth trajectory. If you’re planning to double in size, move to the cloud, or undergo a major system migration in the next two years, your MSP needs to support that. Not all providers are equipped for change — some are only good at steady-state management.

The Questions Most Companies Forget

Who actually does the work? Some MSPs offshore their Level 1 and 2 support. If that matters to you — and for many Australian companies it does, for timezone, compliance, or communication quality reasons — ask explicitly.

What’s your staff turnover? High turnover in an MSP means your environment is constantly being learned by new people. That’s expensive in terms of incident handling quality and resolution times. A good MSP will be transparent about retention.

Can I talk to your escalation engineers? The people you interact with during sales are not the people who’ll be working on your account. Ask to meet the actual technical team. If they won’t arrange that, it’s a red flag.

What does the transition look like? Moving to a new MSP is operationally disruptive. A good provider has a documented transition process with clear milestones, risk mitigation, and parallel running periods. If their answer is vague, they haven’t done it enough.

What’s your approach to proactive work? The difference between a good and mediocre MSP often comes down to proactivity. Are they identifying and fixing problems before you notice them, or are they just responding to tickets? Ask for specific examples of proactive improvements they’ve made for existing clients.

SLAs: Read the Fine Print

SLAs are the contractual backbone of an MSP relationship, and they’re also where most companies get caught out.

Response time versus resolution time. An SLA that promises a 15-minute response time sounds great until you realise “response” means “someone acknowledged the ticket.” Resolution — actually fixing the problem — might have a completely different (and much longer) target.

Priority classification. Who decides whether an incident is Priority 1, 2, or 3? If the MSP controls classification, they have an incentive to downgrade incidents to meet their SLAs. Make sure there’s a clear, objective classification matrix and that you retain the ability to escalate.

Exclusions. Read the exclusions carefully. Many SLAs exclude third-party service outages, planned maintenance windows, and incidents caused by changes you requested. Depending on how broadly these exclusions are defined, they can effectively gut the SLA.

Penalty mechanisms. What happens when the MSP misses their SLA targets? Service credits are common but often capped at a percentage of the monthly fee. If your business loses $50,000 during an outage and the service credit is $500, the incentive alignment is broken.

Reference Checks: Do Them Properly

Every MSP will give you references, and those references will say positive things. That’s expected — nobody provides bad references voluntarily. To get useful information:

Ask for references of similar size and complexity. An MSP that’s excellent at managing infrastructure for a 50-person company might struggle with a 500-person one, and vice versa.

Ask about the transition period. The first three to six months with a new MSP are the hardest. How did it go? What surprised the client? What would they do differently?

Ask what hasn’t gone well. Phrase it gently, but push for honesty. Every relationship has friction. The nature of that friction tells you a lot.

Ask if they’d choose the same provider again. This question cuts through polite endorsements. A pause before answering is informative.

Contract Structure Matters

Avoid long lock-in periods without performance clauses. A three-year contract is reasonable, but it should include performance benchmarks and exit provisions if those benchmarks aren’t met consistently.

Be wary of low base fees with expensive change requests. Some MSPs deliberately price their base contract low and then charge significant fees for any change — new user setups, firewall rule changes, access modifications. Those costs add up quickly. Ask for a schedule of common change request fees before signing.

Understand the exit provisions. What happens when the contract ends? Who owns the documentation? What data do you get back? How much notice is required? How does knowledge transfer work? These details matter enormously when you eventually transition away.

The Incumbent Advantage

If you already have an MSP, switching has a real cost — operational disruption, knowledge loss, and a learning curve for the new provider. Factor that into your evaluation. A marginally better provider might not justify the transition cost. A significantly better one probably does.

Evaluate your incumbent on the same criteria you’d apply to a new provider. Sometimes the exercise reveals that your current MSP is actually decent and the problems are more about communication or scope definition than capability.

Final Thought

Choosing an MSP is one of the most consequential IT decisions a mid-market company makes. It’s a relationship, not a transaction, and treating it like the latter almost always leads to disappointment. Invest the time in proper evaluation, ask the uncomfortable questions upfront, and make sure the contract protects you when things don’t go as planned. Because at some point, they won’t.