The 90-Day IT Contract Audit: What to Do Before Your MSP Arrives Ready to Negotiate

A few months ago, I sat with a client who had eight weeks until contract renewal. He knew the pricing was wrong and had a rough sense of what he was overpaying, but he did not have the data to prove it, and his supplier knew it.

We got him most of the way there in time, but I would not want to do that again.

Ninety days is not long, especially when you consider that most managed service deals run three to five years. But it is enough, if you start at 90 days rather than eight weeks. This article is a practical framework for how to use that time, and it builds on the previous two in this series: The first on building a cost-enriched CMDB as a foundation for commercial control, and the second on what AI is actually doing to MSP delivery costs at renewal.

The clock is already running

Your MSP has been preparing for this renewal for months. They know the contract end date, the margin they want to protect, and what it would cost you to switch. Their account team has a position. In most cases, yours does not.

Ninety days gives you time to build a real position. Wait until 60 or less, and you’re negotiating on their clock, not yours.

Three years ago, this was a simpler conversation. Now AI is running across your MSP’s delivery operations, automating service desks, monitoring infrastructure, and managing incidents before they escalate. Their costs are falling. Those savings rarely show up in what you pay. Knowing that, and being able to prove it, is worth more than almost anything else you bring to renewal.

Days 1 to 30: Start with what you know

Start internally, before you look at anything external. You cannot compare your contract against the market without first knowing what you are consuming and what you are being charged for it.

A cost-enriched configuration management database, one that is accurate, continuously updated, and linked to actual consumption, lets you cross-reference MSP billing against real usage. Most organizations find the gap is larger than expected. The consumption audit almost always delivers more savings than the price negotiation. The €48,000 I described in my first article came entirely from removing unused infrastructure before any rate conversation had started.

Pull the rate card history, the last 12 months of true-up invoices, and the actual consumption volume reports by tower. These three documents will tell you more about where you actually stand than any conversation with your account manager.

Break the contract into service towers: e.g., Workplace, Service Desk, Server, Storage, Cloud Operations, Network, Security, SIAM (Service Integration and Management). For each one, record the unit pricing. Cost per device, cost per user, cost per ticket. Without this, any external comparison is matching your actual costs against someone else’s estimates.

Days 31 to 60: Find out what the market actually charges

Your finance and procurement teams almost certainly do not have live pricing data across multiple service towers and geographies. Your MSP does. That gap only closes one way.

I’ve seen buyers spend the full 90 days polishing the internal audit and never get to this step. That’s a mistake too. The audit tells you what you’re paying. It doesn’t tell you what you should be paying.

An independent benchmarking advisor like IDC, who works across many similar engagements, has that data. The benchmark needs to be honest and accurate, using peer contract pricing from contracts signed in the last 12-18 months. A large multinational contract prices differently from a regional one, even for identical services. Scale, geography, scope, and contract length all matter. Any comparison that skips those variables is not a benchmark. You have a guess dressed up as one.

Where is AI showing up in the price? Most buyers have not asked. Across Service Desk, incident monitoring, and routine infrastructure management, it is driving down MSP delivery costs. The service desk is almost always the most negotiable tower, and it is also where AI has the greatest impact on delivery costs. Your MSP is not likely to point that overlap out to you.

Days 61 to 90: Build the case

Not every finding is worth pursuing. Focus on the towers where your price is furthest above market and your spend is highest. A 20% gap on a small tower matters far less than a 10% gap on your largest.

Evidence moves suppliers. A vague ask for a better rate does not.

A concise summary of where your pricing sits against the market, supported by independent benchmarking data, opens a different kind of conversation. And while you have that conversation, push on contract structure too.

When AI agents handle the delivery, pricing per input, per ticket, or per device starts to misrepresent both the cost and the value of what you are buying. Outcome-based models price on what the service actually delivers: resolution rates, uptime, incident reduction. Renewal is the right time to push for this on at least one or two towers. But define the outcomes carefully. A poorly written outcome-based clause can end up working in the MSP’s favor just as easily as yours.

Your MSP arrived ready. Did you?

Your MSP has a position, prepared in advance, supported by data, and reviewed by their account team. Most buyers arrive with goodwill and a vague expectation that a long-standing relationship will deliver a fair price, but it often does not. The clients that consistently walk away with fair contracts have two things most buyers don’t: an independent market price benchmark, and a consumption audit that’s actually current. Everything else in this article, the AI economics, the outcome-based pricing, the tower-by-tower comparison, only works once those two are in place.

Ninety days is not a strategy. It is the minimum. Start before your MSP has already decided what you are going to pay.