I have spent 20 years building service delivery teams, and I have watched a lot of mid-market companies sign managed services contracts they came to regret. The pattern is almost always the same. The provider sounded great in the sales cycle. The contract was vague in the places that matter most. And six months in, the company realized the gap between what they bought and what they actually got.
This is not a generic checklist. It is the questions I would ask if I were buying managed IT services for my own company, drawn from running service operations for organizations across Minnesota and beyond.
The Red Flags I Look for First
Before I get to the right questions, here are the patterns that should make you walk away.
24/7 support that is really business hours plus voicemail. Many providers list “24/7 support” on their website. Fewer can tell you, with a straight face, who specifically picks up the phone at 2 a.m. on a Sunday and where that person is sitting. If the answer is “we have an on-call rotation,” ask how many tickets that on-call engineer is handling at once and what the response SLA is for a sev-1 outage outside business hours. The answers will tell you whether their 24/7 is real or aspirational.
Contracts that lock you in but do not define SLAs. A real managed services agreement defines specific service levels: response time by ticket priority, resolution targets, uptime commitments for managed infrastructure, and what happens financially if those targets are missed. If the contract has a 36-month term but the SLA section reads like a marketing brochure, the term is the only thing that matters. The service level is whatever they decide it is on any given day.
“We use ServiceNow” versus “we built our practice on ServiceNow.” Tooling answers are easy to fake. Plenty of providers have a ServiceNow instance that nobody on staff knows how to configure beyond out-of-the-box workflows. Ask to see the platform. Ask how long their senior engineers have been working on it. Ask whether they have ServiceNow-certified architects on staff or whether they are reselling someone else’s expertise. The same test applies to Microsoft. There is a meaningful difference between a provider who is a Microsoft partner because they completed the application form and one who has earned the Microsoft Solutions Partner designations for Modern Work, Security, and Copilot.
Onboarding that starts with a discovery call after the contract is signed. If the provider does not understand your environment before the contract, they will not be able to support it after the contract. A serious provider runs discovery during the sales cycle, not after, and walks into the contract knowing what your stack looks like, where your gaps are, and what the first 90 days will focus on.
Questions That Actually Reveal How They Operate
Once you have screened out the obvious red flags, you need to test how the provider runs day to day. These are the questions that separate the providers who can talk a good game from the ones who can deliver one.
“Walk me through your last sev-1 incident from detection to resolution.” This question is hard to fake. Listen for specifics: how the alert came in, who acknowledged it, what the escalation path looked like, how communication flowed to the client, and how long the post-incident review took. If the answer is rehearsed marketing language, you are talking to a sales team, not an operations team.
“Show me your ticket dashboard.” A real provider runs their service desk on a platform that captures every ticket, categorized, with response and resolution times. They should be able to show you the dashboard live. If they cannot, or they hesitate, ask why. The answer will tell you whether they actually measure the things they claim to deliver.
“What does your offboarding process look like?” I know this sounds backwards in a sales conversation. It is the most important question you can ask. A provider who is confident in their service is happy to tell you exactly how a transition out works: data export, knowledge transfer, runbook handoff, timeline. A provider who hedges is telling you that lock-in is part of their business model.
“Who specifically will be on my account?” Big providers love to talk about their bench depth. What you actually need to know is who picks up your calls, who runs your monthly business reviews, and who shows up when something breaks. Ask for names. Ask for their tenure. Ask whether they are dedicated to your account or shared across many.
A Scoring Framework You Can Actually Use
When you have narrowed your list to two or three providers, score them across the dimensions that determine whether the relationship will work. Weight the categories based on what matters most for your business.
| Category | What to Evaluate | Weight |
|---|---|---|
| Coverage model | Real 24/7 staffing, escalation paths, geographic redundancy | High |
| Technology depth | Platform certifications, named senior engineers, proof of expertise | High |
| SLA definition | Specific response and resolution targets by priority, financial remedies | High |
| Onboarding plan | Discovery completed pre-contract, 30/60/90 plan, named transition lead | Medium |
| Communication cadence | Monthly business reviews, quarterly strategy reviews, named account owner | Medium |
| Tooling transparency | Ticket platform access, reporting cadence, real-time visibility | Medium |
| References | Three references in your industry or company size, recent contracts | High |
| Cultural fit | Response style, willingness to push back, advisory posture | Medium |
| Offboarding clarity | Documented exit process, data ownership, knowledge transfer | Low to Medium |
For each provider, score 1 to 5 in each category, multiply by the weight, and total. The highest score is rarely the cheapest provider. It is rarely the largest. It is the one that runs their operation in a way you can actually evaluate.
What to Verify Before You Sign
Before any contract gets signed, three things should be confirmed in writing or in an actual demonstration, not in a sales deck.
First, the SLA section should specify response and resolution targets by ticket priority, what counts as priority 1 versus priority 3, and what financial credits or remedies apply if the targets are missed. If the contract is silent on remedies, the SLA is a target, not a commitment.
Second, the partner certifications should be verified directly. The Microsoft Partner Center directory shows current designations and locations. ServiceNow’s partner finder does the same. If a provider claims a designation they cannot demonstrate in a public partner directory, that is a flag.
Third, you should talk to at least two reference clients who left the provider, not only the ones who stayed. Stable clients tell you what is working. Departed clients tell you what broke and how the provider handled it. Both perspectives matter.
The Virteva Model
I am not going to pretend this is an objective post. I run service delivery at Virteva, and the framework above describes how we operate because it is how I have always believed managed services should work.
Our service desk runs on ServiceNow, and our practice was built around the platform from the start. Our 24/7 coverage is a hybrid model with engineers in Minneapolis and a delivery center in Manila, which means the engineer who picks up at 2 a.m. is not someone pulled out of bed for an on-call rotation. They are mid-shift. Our Microsoft Solutions Partner status covers Modern Work, Security, and Copilot, and we are a ServiceNow Elite Partner. Our NPS is 87 across active clients, and our average client tenure is over five years.
If you are evaluating providers and want a candid conversation about how to run the process, reach out. I am happy to share the scoring framework and walk through what discovery should look like, even if you decide we are not the right fit. The goal is helping you make a good decision. The wrong managed services provider costs more than money. It costs the time and trust you spend trying to make a bad relationship work.
