The first four weeks of someone new working remotely, week by week

A remote hire is almost never decided in the selection. It is decided in the first four weeks, and it is decided by small things: an access that did not work on day one, a policy nobody had written down, four people giving different instructions.

Here is what has to happen each week for there to be autonomy by the fourth, and exactly what breaks when a piece is missing. It works as a list: you can tick it off.

Before day 1

What gets prepared beforehand decides the rest of the month. Five things, and all five are conditions rather than recommendations.

• Accesses created and tested. Created is not enough: tested. Someone logs in with those credentials before the first day and checks that the real work can be done, not just that the login works.
• The policy in writing. What is done with the usual, and what is done with the odd case. One page will do. What will not do is «they will pick it up as they go».
• The canned replies you already use. No need to write new ones: gather the ones that exist, even if they are spread between the helpdesk, your inbox and two people's heads.
• A single point of contact. One person to ask. Others can be involved, but questions come in through one door.
• The two or three tasks for the first month, chosen. Not the whole role: what has to be working on day 30.

What breaks if something is missing. If the accesses are not ready, the first week goes on IT support and the person starts the month feeling like an obstacle. If there is no written policy, every case becomes a question, and by the third question in a row whoever delegated starts to think they have hired more work than they had. If there are four points of contact, there will be four sets of criteria, and the person will end up following whoever spoke last.

Choosing well who you bring in is the other half of this, and it comes earlier: what gets checked before anyone is put in front of you is in how we choose the person who will work with you.

Week 1: training, and reviewing everything that goes out

Week 1 is not for producing. It is for the person to understand your product, your customer and your criteria, and for you to see how they work.

Three short sessions are enough, and three is better than one four-hour block: product, tools and policy. Between sessions, genuinely real work — cases that exist, not exercises — and everything that goes out gets reviewed before it goes out. Case by case, no exceptions, all week.

That upfront review looks expensive and it is the cheapest thing in the month. Every correction in week 1 prevents the same mistake repeated fifty times in week 3, and above all it teaches the criteria far better than any document: the person does not learn the rule, they learn why the rule.

Two signs the week is going well: the questions get more specific, and they start flagging things you had not seen.

What breaks if it is missing. Without upfront review, the first week produces work that has to be undone, and undone in front of the customer. Without training, the person fills the gaps with whatever seems reasonable — and reasonable, without context, almost never matches yours.

Weeks 2 and 3: review by sampling

Here the method changes: stop reviewing everything and review a sample. A handful of cases a day at first, a handful a week by the end, picked without warning and mixing the easy with the awkward.

Sampling does two things at once. It frees up the time of whoever delegated — which was the point — and it keeps the signal: if something is going crooked, it shows up in the sample. What must not happen is jumping from reviewing 100% to reviewing nothing, which is the jump nearly everyone makes and the reason month two brings surprises.

These two weeks are also the moment to hand out permissions with a limit. Not «they can issue refunds», but «they can issue refunds up to this amount, and above it they escalate». A permission with a written limit removes half the escalations made out of caution without opening any risk.

And it is when the process document really starts to fill up, because the new person is the one who finds the missing cases. Let them add those the same day, without asking permission, and validate them afterwards.

What breaks if it is missing. If everything is still reviewed in week 3, there is no delegation: there is a permanent trainee and two people doing one person's job. If nothing is reviewed, the first big mistake is discovered through a customer complaint, which is the most expensive way to discover it.

Week 4: autonomy around 80%

The goal of the fourth week is not for the person to do everything. It is for them to do most of it alone and to recognise the rest. That rough eighty per cent is the realistic number: the remaining twenty are the cases that should be escalated, because they involve a decision that is not theirs to make.

It is also the week of the first performance review, and it is worth having data in front of you: volume handled, times, escalations and what has repeated. What each of those numbers measures, what distorts it and what it has to be read alongside is developed in the article on what to measure in customer support.

The useful version of that conversation has three parts and takes half an hour: what is going well and stays as it is, what changes this week, and what they need from you that they still do not have. The third is the one almost nobody asks and the one that usually explains the most.

What breaks if it is missing. Without a first review, week 4 looks like week 3, and so does month 2. Autonomy does not arrive with the passing of time: it arrives because somebody said, with cases in front of them, how far the person decides alone.

Month 2 onwards

What sustains a role is not the start, it is the follow-up. A monthly review with the person, the sampling reduced but alive, and the process document updated whenever a new case appears. Without that, a role that worked in month one degrades by month six without anyone being able to point to when.

In our case, that follow-up is not optional and does not depend on the client asking for it: performance and duties are reviewed every month against the objectives that were set. The full split of what you do and what we do, with these same four weeks summarised, is in how a team works.

The list, in short

• Before day 1: accesses tested, policy written, canned replies gathered, one point of contact, two or three tasks chosen.
• Week 1: three training sessions, real work, upfront review of everything that goes out.
• Weeks 2 and 3: review by sampling, permissions with limits, the process being written as you go.
• Week 4: autonomy around 80%, first performance review with data, agreement on what changes.
• Month 2: monthly review, live sampling, document kept current.

None of this is complicated. What happens is that it gets improvised, and what is improvised in week 1 gets paid for in month 3. If you are bringing someone into a customer support role or into any other delegable job, prepare the week before it starts before you prepare the first one.