Your IT contract promises a one-hour response. Nobody has ever checked

Somewhere in your IT support agreement there is a number. One hour, four hours, next business day. It was probably one of the things you compared when you chose the provider, and it is almost certainly the thing you have never measured.

Every article on this subject tells you that a response is not a resolution. That is true, useful, and written mostly by the firms who sell the contracts. It is also the easy half. The harder half is that the number is not a fact about the provider. It is a setting, and they chose it.

One number, five settings behind it

What starts the clock. A request can reach a provider three ways, and each one starts the timer somewhere different. A portal ticket is stamped the second it is submitted. An email to a support address starts when the system ingests it, or when a human reads it, depending on how the queue is wired. A phone call starts whenever the person who answered gets round to logging it. Same problem, three start times, potentially hours apart. The channel your staff actually use is usually the phone.

Whether the hours are working or calendar. A four-hour target on working hours is a different promise from four hours flat. Report something at half past four and one version answers before six; the other answers tomorrow morning and is still inside its target.

Whose working hours. If your people start at seven and the support desk starts at nine, the first two hours of your day are outside the agreement. That may be perfectly reasonable. It should not be a surprise.

Who decides the severity. This is the largest lever in the whole document, and the one nobody reads. Severity picks which target applies, and in most agreements the provider assigns it. A P1 becomes a P3 and the four-hour promise becomes next week without anybody breaking a term. Ask what happens when you disagree with the grading, and ask before you need to.

What a response is permitted to be. In a fair number of contracts an automated acknowledgement satisfies it. If an auto-reply stops the clock, the number in your contract measures the mail server.

A provider who assigns the severity is grading their own performance.

I know where the settings are because we built one

We wrote a ticketing system for a payroll partner, to replace an off-the-shelf tool that could not keep up with how their week actually worked. It has an escalation ladder that changes by time of day and day of week, and ageing counted on working days only.

Every one of those rules had to be decided on purpose, and every one of them has a version that flatters whoever is being measured. "Working days only" is the honest reading of a rule that could just as easily have been written to make Friday afternoon vanish. Where the ladder steps up at four o'clock rather than five is a judgement about the customer's evening, not the desk's. None of it is dishonest. All of it is discretionary, and the discretion sat with us.

That is the part a generic explainer cannot tell you, and it is the reason to stop reading the number and start reading the definitions.

The audit nobody needs their provider's cooperation for

You do not need a report, a meeting or a new tool. You need your own sent mail and about an hour.

01Pick a quarter and search your sent items for everything you reported to support
02For each one, write down when you asked and when a human being actually answered
03Note how it was reported: portal, email or phone
04Put the middle of that spread next to the number in your contract

Do not build a legal case. You are answering one question: does the number describe your experience?

Two things can come out of it. The times broadly match the contract, in which case you have spent an hour buying the right to stop thinking about it, and there are better places for your attention. Or there is a gap, and the gap is almost never explained by anyone being lazy. It is explained by one of the five settings above, which is a much more useful thing to know, because a setting can be renegotiated and a grievance cannot.

Your sent mail will under-count the phone calls, and that is a finding rather than a flaw in the method. If most of what your firm reports goes in by phone, and the phone is the channel with the vaguest start time, you have just learned which part of the agreement is doing the least work for you.

At renewal, ask for the two things that cannot be configured away

Not a smaller number. A one-hour target with no definitions is worth less than a four-hour target with clear ones, and pushing the number down is the change a provider can concede most cheaply.

Ask instead for the definitions in writing: what starts the clock on each channel, which hours it runs on, who assigns severity and what happens when you dispute it, and whether an automated acknowledgement stops it.

Then ask for a monthly report you did not have to request. That one is the real test. Any provider can produce a report when a customer is annoyed. A provider whose report arrives every month unprompted, with the misses still in it, is one whose system was built to be looked at. It is also, in our experience of building the thing that produces those reports, the single hardest commitment to make and the most telling one to keep.

We are in the odd position of having built the clocks and of being a supplier these questions get put to. We would rather be asked. The security version of this conversation is four questions and one email, and it belongs in the same renewal; the first hour of any new relationship should have covered some of this and usually does not.

If you have two contracts on the desk and cannot tell which is better, the numbers are not where the difference is. Bring us both and we will read the definitions.

We work with businesses across Knutsford, Alderley Edge, Wilmslow, Altrincham, Stockport and Warrington, and remotely for clients anywhere in the UK.