Hiring remotely on your own or with someone behind it: what actually fails

Hiring someone remote is easy. You post the role, applications come in, you speak to three, you pick one and they start on Monday. The easy part has been solved for years.

The hard part is it still working in month three. And when it does not work, it is almost never because the person was bad: it is because nobody was holding up the five things that have to be held up for a remote role to stay standing.

This article is the list of those five things. It is useful for deciding whether you do them or whether you would rather someone else did, and it is just as useful if you decide to do them yourself: at least you will know what falls to you. It holds for any continuous operational role, from customer support to admin, catalogue or data.

Selection with no practical test

The first leak happens before anyone starts. An interview measures how well someone explains themselves, not how they work. In an operational role that is not enough, because what you need to know is whether they can do the job in the tool where they will be doing it.

Doing it on your own, the practical test is the first thing to go. Preparing a realistic case takes a while, marking it takes another, and multiplying it by six candidates is an entire afternoon you do not have that week. So the decision is made on the conversation and on the CV, and the CV says they have used your helpdesk without saying whether they used it well.

Then there are two checks almost nobody does that explain half the nasty surprises: speaking to whoever directed their work before —not the reference they choose, but the person who asked them for things— and checking the connection and the workspace before day one. The candidate who says all the right things in the interview and cannot connect on day 3 does exist, and always turns up because nobody asked.

We have written ours out step by step in how we choose the person, with the four checks we never skip. Use it as a shopping list, even if you hire on your own.

The onboarding nobody has time for

The second leak is the most common and the quietest. You hire someone because you are swamped. That person arrives on Monday and needs two weeks of your time exactly when you do not have two hours. So you give them the access, a couple of explanations and your trust.

What happens next is predictable: they work by guessing. They answer what they think you would answer, apply the policy they deduced from reading five old cases, and do not ask because it feels like they should not be a nuisance. By the time the first serious mistake shows up, it has been repeating for weeks.

The antidote is not a fifty-page manual. It is three things in writing, even if each is one page: what gets answered and what gets escalated, what the policy is when the customer is right and the rule says no, and who to ask when there is no answer. With that, the first week stops being guesswork. Without it, it does not matter who you hired.

And there is one scheduling detail that helps more than it sounds: in the first week the work gets reviewed before it goes out. That is not mistrust, it is the only way to correct a judgement call before it becomes a habit. From there autonomy rises on its own, and it rises faster the more written down your policies are.

The equipment, the connection and where your data goes

The third is the one least thought about at the start and the hardest to fix later. If the worker provides the equipment, the equipment is whatever there is: the laptop they already had, their home connection, their antivirus, their other accounts left open. When something is slow or breaks, there is nobody to call, and the role stops until it is sorted.

And there is a more uncomfortable question: where your data is while somebody works for you. If that person downloads a customer list onto their computer, that list is on a computer you do not control. The day they leave, you do not control what stays there either.

Our answer is for the work to happen inside a virtual desktop that we manage: the person logs in there, their access is in their own name and is revoked the same day, and the data does not cross over to their personal machine. It is explained with the detail it deserves in where your data is. If you hire on your own, the question is still yours and it is worth answering before day one, not after the first incident.

The absence nobody covers

The fourth is always discovered on the same day: the Tuesday that person falls ill. An operational role covered by one person with no backup is a role that disappears when they do. Tickets pile up, returns stop, and the work you had delegated comes back to your desk with two days of delay on top.

Holidays are the predictable version of the same problem, and they still catch everyone out. Nobody plans who answers in August until August arrives.

This is not fixed by hiring better. It is fixed by deciding in advance what happens on those days: who covers, with what priority, and what can wait. In our case it sits inside the fee —20 days a year of managed absence, with no paperwork on your side— and you can read it alongside the rest in what happens if something goes wrong. On your own it is a spreadsheet and an awkward conversation, but you have to have it all the same.

Month three: the person who switches off

The fifth breaks nothing suddenly and is the one that takes down the most remote hires. The first two months go well: there is novelty, there is drive, there are questions. In the third the pace starts to drop. Replies get shorter, initiative disappears, the work gets done and nothing more.

It is almost always the same thing: that person has nobody to talk to. They work alone, they are in no team conversation, nobody tells them whether they are doing well and nobody tells them what is happening at the company they work for. A remote role without follow-up does not rebel: it switches off. And it does not matter which role we are talking about: it happens the same in customer support as it does in admin.

What prevents it is boring and it works. A fixed conversation with someone who looks at their work, written objectives that get reviewed, and someone to take a problem to who is not you —because you do not take problems with a client to that same client. That is exactly the part a provider can hold up and that you, in the middle of your week, will not hold up every week.

So: on your own, or with someone behind it?

On your own makes sense if the role is simple, the person works with data you are not worried about, and you have real time to direct and support them. It is cheaper, you move faster, and it works more often than those selling the opposite will say.

With someone behind it makes sense when the role is continuous, when what is being handled is your customers' data, and when the cost of the role disappearing for two days is high. You are not buying the person: you are buying that the selection was done properly, that the equipment and the environment are sorted, that absence is covered and that somebody tracks their performance while you are looking elsewhere.

The honest way to decide is to go through the five things above and mark which ones you will genuinely do —not which ones you should. If you mark three or fewer, the role will fall over in month three, and it will not be the fault of whoever you hired.