“We respond within 30 minutes” is a phrase you will find in almost every IT support provider’s proposal. The problem is that it can mean anything: from an automatic email saying “your request has been received” to a specialist who is sitting at your computer half an hour later and working on the problem.
For a CEO or a finance director the difference only becomes visible when the accounting system is down on month-end closing day and the helpdesk has “responded” with a ticket number and silence. Then it turns out that the proposal promised a response time, and what you needed was a resolution time.
In this article we explain how these two differ, how a priority system should work, what exactly has to happen in the first 30 minutes, how to measure it, and which questions to ask a provider before you sign.
Response time and resolution time are two different commitments
Response time is the time from receiving a request to the moment a live specialist starts working on it. Not an automatic email, not a number in a system, but a person who has contacted you, understood the problem and started acting.
Resolution time is the time until the employee can work again. It depends on how complex the problem is, which is why an honest provider never guarantees it as a single number for all requests. A stuck Outlook can be fixed in ten minutes, while a failed server disk takes several hours even when everything is done perfectly.
A good service level agreement (SLA) defines both: a firm response time by priority and a target resolution time by type of problem. If you see only one number in a proposal, ask which one it is.
There is also a third time that is talked about less often: the workaround time. When a printer breaks, what matters most to the employee is being able to print now, not knowing when a part will arrive. A good helpdesk first redirects to another printer and only then orders the repair.
Priorities: not all requests are equal
If a helpdesk handles all requests in the order they arrive, a critical problem waits behind a request to install a new font. That is why requests are sorted by impact. In our practice we use a four-level system:
- Critical: the whole company or an essential process is down, such as a server, the accounting system, email for everyone, or the internet connection at the office. A suspected security incident always lands here.
- High: one department cannot work, or one employee cannot work at all, such as a broken computer, a blocked account, or a VPN that does not work for a remote employee.
- Medium: the employee can work, but with disruption, such as a slow computer, one printer that is not working, or an application that sometimes freezes.
- Low: requests and improvements, such as a new application, access to a shared folder, or a how-to question.
The priority is set by the helpdesk, not by the author of the request. That matters because to everyone their own problem looks the most important. But a good helpdesk asks not only “what is not working” but “how many people cannot work, and what happens if this stays down for another hour”. That question separates a critical problem from an annoying one in a minute.
A 30-minute response should apply at least to critical and high priorities. For low priority, the next business day is fine. What matters is that the employee knows this immediately and does not have to guess.
What happens in the first 30 minutes
A good response is not just speed. It is a sequence of specific actions. Here is what it looks like on our side when a high-priority request arrives:
- First five minutes: the request is registered, a priority is assigned, and we determine whether it is an isolated problem or whether several employees have already reported the same thing. If monitoring is showing a server or network alert at that moment, the requests are linked.
- By minute ten: a specialist contacts the employee through the channel they used and confirms that the problem has been understood correctly. This is often where “the internet is down” turns out to mean “one website will not open”.
- By minute twenty: the specialist connects remotely or, if it is a server problem, starts diagnostics. If the fix is clear, it is applied immediately. If not, the employee is offered a temporary way to keep working.
- By minute thirty: the employee knows what is happening and when to expect a fix. If the problem is critical and affects the whole company, the responsible person on your side is informed, not just the author of the request.
Notice that the problem is not necessarily solved within these 30 minutes. But it is understood, it has a responsible person, a temporary solution and clear communication. That is what lets the employee and the manager plan instead of worrying.
A badly run helpdesk sends an automatic confirmation in the same 30 minutes and calls two hours later asking “so what is not working?”.
How to measure whether the helpdesk keeps its promises
Commitments without measurement are only words. A good provider produces a monthly report on its own, and you should be able to read it. What to look for:
- Number of requests by priority and type. If half of the requests are about the same printer, the problem is not the helpdesk, it is the printer.
- Response time compliance as a percentage by priority. The target should be close to 100 for critical and high.
- Average and median resolution time. The median matters more, because one long repair distorts the average.
- Share of reopened requests. If the same problem comes back, it was papered over, not solved.
- Share of requests resolved on first contact. This is the best indicator of helpdesk competence.
- Employee ratings after a request is closed, if the provider collects them.
It is important that time is measured from the moment the employee reached out, not from the moment the provider registered the request. Otherwise the 30 minutes begin when it suits the provider.
We recommend reviewing not only the numbers with your provider once a quarter, but also the recurring causes. Prevention is born from this review: a replaced disk, an updated application, trained employees. This is how our IT support for businesses works: the helpdesk is not only about fixing faults, it is also the source from which we see what needs to change.
What to ask a provider before you sign
In a conversation with a prospective provider, these questions show within minutes whether “30 minutes” is a commitment or a slogan:
- From what moment is response time counted, and what counts as a response?
- Does response time differ by priority, and who sets the priority?
- Through which channels can employees contact the helpdesk, and are all channels equally fast?
- What happens with critical problems outside business hours and on weekends?
- How many specialists will work with our company, and who covers when they are on holiday?
- What report will we receive every month, and can we see an anonymised one now?
- What happens if the response time is missed, and are consequences written into the contract?
- Does the helpdesk have access to the servers and the Microsoft 365 tenant, or is that a separate service?
If a provider answers “it depends” to at least half of these questions, you will most likely get a ticket number and silence.
Where to start
If your helpdesk today works on the principle of “call Tom”, we suggest these steps:
- For one week, record every IT problem in the company: who reported it, when, when it was responded to and when it was resolved. This table alone usually shows how much working time is lost waiting.
- Define what counts as a critical problem in your company. The accounting system on month-end closing day and email for the sales team will certainly make the list.
- Ask your provider for a written SLA with response time by priority and a measurement procedure. Our service level description is public, so you can use it as a basis for comparison.
- Make sure employees know where to turn. The best helpdesk does not help if people keep calling a colleague.
- After the first month, review the report and check whether the numbers match what employees experienced.
A response within 30 minutes is achievable and it is not a luxury. But it means a process, not a promise.
