Nobody reads their IT support contract until something has gone wrong, and by then the contract is the referee, not the negotiation. The document decides how fast “urgent” legally is, which of your problems cost extra, what happens when you want to leave, and whether that missed renewal window just bought another year. It is a short document with long consequences, and every clause in it was written by the provider’s side of the table.

This guide walks through the contract piece by piece: what a service level agreement actually commits to, the inclusion and exclusion architecture where the real price lives, the terms that bite at renewal and exit, and a checklist to run over any agreement before signing. Read it before the quote stage and the whole market becomes easier to compare.

The SLA: what “fast response” means in writing

The service level agreement is the contract’s engine room, and it works on two distinctions most buyers miss.

Priority levels. Good SLAs classify issues, typically P1 to P4: P1 for business-down (nobody can work), P2 for major impairment (a team or system down), P3 for single-user problems, P4 for routine requests. Each level carries its own committed timescale, which is what makes “we respond fast” measurable: a P1 response in 30 minutes and a P4 in two days can both be excellent; a single vague promise covering both is neither.

Response versus resolution. “Response” means an engineer engages with the ticket; “resolution” means it is fixed. Providers commit firmly to response times because they control them, and loosely (or not at all) to resolution times because problems vary. That asymmetry is normal, but you should know which one you are reading: a contract promising only “response within 1 hour” has promised an engineer will *start* within an hour, nothing about when you will be working again. The better agreements add resolution targets by priority, even if framed as objectives rather than guarantees.

Teeth. What happens when a commitment is missed? Service credits (a rebate against the monthly fee) are the standard mechanism, and their presence signals a provider who expects to hit the numbers. An SLA with no consequences is a mood board. Ask; the reaction is diagnostic, as our choosing-a-provider questions cover.

Measurement hours. A “4-hour response” measured in business hours means a 4:55pm Friday ticket can be legitimately touched at 12:55pm Monday. Neither wrong nor hidden, but worth understanding against your operating pattern, and it is where out-of-hours arrangements attach if you need them.

Inclusions and exclusions: where the price actually lives

Two contracts at £55 per user can be £15 apart in real cost, and the difference is the scope architecture:

Typically included in a fully managed agreement: help desk, monitoring, patching, security tooling management, backup oversight, user administration, vendor liaison. Typically excluded and billable: projects (migrations, new servers, office moves), new-starter hardware setup beyond a threshold, on-site visits beyond an allowance, out-of-hours work, and third-party costs (licences, hardware, courier-grade problems). Exclusions are legitimate; invisible exclusions are not. The contract should let you predict, for any plausible event, whether it is covered, and the 36-month cost modelling that makes quotes comparable depends on exactly this.

Watch two specific patterns. The fair-use trap: unlimited support “subject to fair use” with fair use undefined, giving the provider a discretionary lever. Ask for the definition in numbers. The project boundary: where does a big support ticket end and a billable project begin? Better contracts define it (hours, or a named threshold) so the classification is not a negotiation after the fact.

Term, renewal and the calendar trap

Twelve-month terms are standard; 24 and 36 exist and should buy meaningful discount. The clause that actually catches businesses is auto-renewal with a notice window: renewal for a further full term unless notice is given, commonly 60 to 90 days before the end date. Miss the window by a week and the polite version of what happens is nothing good. Diarise the notice date the day you sign, and prefer contracts that renew monthly after the initial term instead of relocking a year.

Price-rise mechanics belong here too: uncapped mid-term increases are a red flag; annual adjustment at renewal, capped or index-linked, is the fair pattern.

Exit terms: the clause you negotiate while you still can

Everything about leaving is decided at signing, when you have leverage, not at leaving, when you do not. The contract should commit the provider to a documented offboarding: handover of all credentials, licences, data and documentation, within a stated period, at a stated (ideally zero, at worst fixed) cost. It should also confirm what our switching guide treats as gospel: licences, domains, tenancies and admin accounts registered to *your* business throughout, so the exit clause governs cooperation, not hostage release.

The fair-contract checklist

Run any agreement past these before signing:

A provider whose standard contract already passes most of this list has told you something valuable before the relationship starts. Ours does, and we walk through it clause by clause during the health-check conversation, because a client who understands the contract complains less and renews more, which is the only incentive alignment that matters.

Frequently asked questions

What should an IT support contract include?

Defined service scope with itemised inclusions and exclusions, an SLA with priority-based response times and consequences, term and renewal mechanics, ownership confirmation for accounts and licences, and documented exit obligations.

What is a good SLA response time for IT support?

For SMEs: P1 business-down response within 15 to 60 minutes, P2 within a few hours, P3 same or next business day, P4 within days, measured against stated hours of cover. The realistic benchmark is less about the numbers than whether they are contractual.

What’s the difference between response time and resolution time?

Response is when work starts; resolution is when the problem is fixed. Contracts commit hard to the first and softly to the second; know which you are reading before comparing providers.

How long should we sign for?

Twelve months initially, moving to rolling monthly after the initial term. Longer commitments are fine when discounted and paired with fair exit terms; auto-renewal into fresh fixed terms is the pattern to avoid.

Can we get out of an IT support contract early?

Only per its termination clause: typically for material breach (with a cure period), sometimes with an early-exit fee. Practically, the notice window and renewal date matter more; most businesses leave at term, and the ones who suffer are those who missed the window.

What are service credits?

Rebates against your fee when SLA commitments are missed: modest money, strong signal. Their real value is that a provider offering them has instrumented their own performance, which correlates with everything else you want.

Read ours before you sign anyone’s

The fastest education in IT contracts is putting two side by side with the checklist above. We are happy to be one of the two: our free IT health check comes with our standard agreement and SLA to take away, annotated against this page. Get in touch and bring the contract you have now; the comparison usually takes twenty minutes and pays for itself at your next renewal window.