Remote support or someone on site: where the line sits

What can be done remotely, and what cannot

The distinction is sharper than people expect, and it lines up almost entirely with one question: do you have to touch something to fix it?

Resolvable remotely is everything that lives in software. An application that will not start, rights that are wrong, a mailbox that is full, a printer that is not found, an update that hung, a password that needs resetting. Most investigation too: reading logs, comparing settings, creating a test user to see whether it is the person or the system.

Not resolvable remotely is everything physical. A laptop that will not power on, a cable that has come loose, a switch that needs replacing, an access point being moved, a screen that has died. And one category that technically could be done remotely but in practice should not be: the first day of someone new, and the moment a whole department moves over to something else at once.

That last one is not technology but people. Someone who has just arrived and sits watching a screen waiting for a mouse to move by itself starts their first day feeling that IT is something that happens to them.

The three models, and who they fit

In practice we see three ways of organising this. They differ not in technology but in who sits where.

• Fully remote. Everything from the service desk, and for physical work you bring someone in per occasion — a local party, a supplier, or one of your own people lending their hands. Fits organisations on one or two sites doing mostly office work, and it is the cheapest model because you do not pay for travel time you do not use.
• Remote with fixed visit days. The same service desk, plus someone who comes by on a fixed rhythm — weekly, fortnightly. That day goes on what has been left: handing out devices, tidying up, and the conversation nobody has over the phone. This is the model that works out best in most estates, and it is chosen least often because it does not fit a quote that only looks at hourly rates.
• Someone on site, permanently. One or more people who are there, with the service desk behind them for work outside their hours and for second line. Fits from around a hundred and fifty workplaces on one site, or sooner where there is production or a front desk and downtime costs money directly.

Which one fits depends on three things: how many sites you have, how much of your work is physical, and what an hour of downtime costs. Not on how big you are.

What a site visit really costs

Amounts do not appear on this site, but the structure of this one does, because that is where the misunderstanding lives.

A visit does not cost the hour someone spends with you. It costs the journey out, the hour, the journey back, and the ticket that landed on someone else during that time. For a twenty-minute fault that has to be fixed on site, you are therefore paying for half a day of attention — and that is not profiteering but arithmetic.

Which is exactly why fixed visit days work out cheaper than visits on demand in most cases. You bundle ten small things into one trip instead of three trips for three things. The price per ticket falls and the sense of availability rises, and those two rarely move together.

What to watch for when comparing quotes: does travel time fall inside the rate or outside it, is there a minimum per visit, and what happens when the engineer arrives and the problem turns out to be something else. That third one rarely appears in a quote and always happens at some point.

Where the agreement comes unstuck in practice

Nearly every SLA promises a response time. Nearly none promises a resolution time, and rightly so, because that depends on what is broken. But it does mean "within four hours" can be something other than you think.

So press on three things before you sign:

• What does response mean? Someone who looks, or someone who starts. An automatic acknowledgement within one minute is also a response.
• Where does the person coming on site travel from? "On site within four hours" is a different promise when the nearest engineer is a two-hour drive away, especially with several sites able to call on them at once.
• What applies outside office hours? With us monitoring and support run 24/7 on a rota, but an on-site visit at night is a separate agreement with practically every provider. Whoever does not make that explicit finds out at the worst moment.

What we put in writing about this sits under IT service desk, and the wider trade-off about what you outsource sits in outsourcing your service desk.

How to work out the balance yourself

You do not have to guess at this. There is a sum you can do with your own tickets, and it takes an afternoon.

Take the tickets from the last quarter and split them into two piles: resolved without anyone visiting, and not. Count the second pile and look at what it is about. In most organisations we take over that is between five and fifteen per cent, and three quarters of it sits in two categories: hardware that has broken and new people.

If you are under ten per cent and it is mostly devices, then fully remote with a local party for the hands is nearly always the answer. If you are above it, look at why — that is more often a sign that something structural is wrong (ageing equipment, a network that is not stable) than that you need more visits. Sending more people does not fix that; it only makes it bearable.

We will do that count with you if you want, and you get it even if you do nothing with us afterwards. It is your data.