TL;DR

Telling people their laptop is being replaced looks like good news from the inside and like an interruption from the outside. The person is not getting a new laptop, they are losing the one they set up exactly how they like it. Say the three things they actually want to know and most of the resistance disappears.

  • They hear disruption, not upgrade. Write the message for that.
  • Three questions every time: when, what happens to my files, what do I lose.
  • Give a date, not a window. A quarter is not an answer.
  • Say what is not transferred. That is the part that causes complaints.
  • Overlap beats handover. Two machines for two days fixes most of it.
  • Announce the programme once, then message each person individually.

Telling people their laptop is changing

Whatever the message says, the reply contains the same three questions. Answer them in the message and you remove most of the thread.

When exactly? People have deadlines. “During Q2” is not an answer, it is a worry with no end date. Give the day.

What happens to my files? Specifically, what moves automatically, what they have to move themselves, and what is simply gone. The third category is the one that generates complaints.

What do I lose? Browser tabs, local tooling, saved passwords, that script in a folder nobody backs up, the particular way their terminal is set up. These are small and they matter, and pretending they do not is why refresh communications go badly.

Why the framing is wrong by default

Internally this is an upgrade with a budget attached. Externally it is somebody taking away a machine that works and that they have spent two years configuring.

Lead with the second reading. “Your laptop is being replaced on the fourteenth, here is what that means for you” lands better than “good news, new hardware”, because the second one is answering a question nobody asked.

What the message should contain

Six lines, in this order. Longer than that and the important part gets skimmed.

Line Content Why it is there
1 The date The first thing they look for
2 What arrives, in one phrase Sets expectations without a spec sheet
3 What moves automatically Removes the main fear
4 What they must move themselves, with a deadline The only action they have
5 How long they keep both machines Turns a cut-over into a transition
6 Who to reply to A name, not a queue

The line most people leave out

Line four. Teams describe what the migration covers and go quiet about what it does not, because the gap is awkward.

Name it anyway. Local files outside the synced folders, anything installed without administrative rights, locally stored credentials, virtual machines, container images. Saying these are not transferred, with a date by which the person should handle them, costs one sentence and prevents the conversation that starts “nobody told me”.

Announce the programme once, then go individual

One company-wide note explaining that a refresh cycle exists, what the interval is, and how people will be told. Then an individual message per person with their own date.

The combined version fails in both directions: a company note with no personal date generates anxiety, and a personal note with no context reads as if something is wrong with their machine specifically.

Give them two machines for two days

Disclosure: RemoAsset, named below, is owned by the same people who publish PeopleOpsHQ. It is described with its limitations stated, on the same terms as every other supplier. This site takes no vendor money.

The single most effective change available here is not a better message. It is overlapping the two devices.

A same-day handover means the person stops work, migrates, discovers something missing, and cannot get it because the old machine has gone. Two days of overlap turns every one of those discoveries into a two minute fix.

It costs you two days of holding one extra device per person, staggered across the programme, which is a small inventory cost against a visible drop in support load and complaints.

What to ask for at the end of the overlap

A confirmation from the person that they have what they need, before collection is booked. One line, and it converts the return from something being taken to something being released.

That matters practically as well as emotionally: the commonest cause of a machine sitting uncollected for months is that the person was not sure they were finished with it, so they quietly kept it.

Once the confirmation exists, the collection should be immediate rather than scheduled vaguely. Collection with packaging and a courier arranged is the mechanism that makes a booked date realistic, and the limitation worth stating is that it cannot generate the trigger. The confirmation step is yours.

Then say what happens to the old one

People ask, and the answer builds more goodwill than anything else in the process. Say whether it is wiped and reissued, resold, or disposed of, and say that the wipe is documented.

Where data protection frameworks such as the consolidated text of the General Data Protection Regulation apply to what was on the machine, a documented erasure against a standard such as NIST Special Publication 800-88 on media sanitization is both the correct handling and the specific reassurance people are asking for when they wonder what happens to their old laptop.

The decision table

Situation Scale Setup Primary pain Recommended starting point
First ever refresh programme Any No precedent Reads as something being taken Company note, then individual dates
Rolling refresh 100 plus Established cycle People do not know their date Publish the date in a place they can look up
Platform change Any Different operating system Tooling and habits lost Longer overlap, named help per team
Engineering teams Any Heavy local setup Environments not transferable Four days overlap, not two
Machine already failing Any Urgent swap No time to prepare Ship first, migrate with support on a call

When somebody says no

Occasionally a person simply does not want the new machine. It is worth understanding why before treating it as resistance, because the reasons are usually practical.

Three come up repeatedly.

Timing. They are mid-project and cannot lose two days. This is almost always reasonable and almost always solvable by moving the date, so move it rather than arguing.

A specific dependency. Something on the old machine that nobody can reproduce, often an old tool, a licence tied to hardware, or a build environment assembled over years. This is a real risk and needs a technical conversation rather than a policy one.

The new machine is worse for their work. Sometimes true, particularly where a standard specification was set for the majority. Worth hearing, because it is the same signal as an upgrade request arriving from a whole team.

Where to hold the line

On machines past end of support. A device no longer receiving security updates is a different conversation, and the right framing is that it is being retired rather than upgraded.

Say that plainly, give a date, and offer whatever help makes the dependency problem solvable. For the backlog of machines in exactly that position, our guide to answering an upgrade request covers the opposite direction of the same conversation, and the two policies should agree with each other.

Record the refusals

Not to police them. Because a pattern of refusals from one team is information about the standard specification, and three people with the same irreproducible dependency is a documentation gap worth closing before the next cycle.

One thing to measure afterwards

How many days elapsed between the new machine arriving and the old one being collected. Not satisfaction, which nobody answers honestly, and not ticket volume, which measures your support team rather than the process.

That one number tells you whether the overlap length is right, whether the confirmation step is happening, and whether people felt finished. When it starts stretching past a fortnight, something in the message stopped being clear, and it is usually line four.

See the collection step that follows a confirmed handover at device retrieval with packaging arranged, or read our guide to answering an upgrade request.

Frequently asked questions

What do people actually want to know?

Three things, in every reply you will get. When exactly, because they have deadlines and a quarter is not an answer. What happens to their files, split into what moves automatically, what they must move themselves and what is simply gone. And what they lose, meaning the saved passwords, local tooling and particular setup that took two years to get right.

How should the message be framed?

As a change happening to them rather than as a gift. Internally a refresh is an upgrade with a budget attached. Externally it is somebody removing a working machine they have configured carefully. Leading with the date and what it means for them lands better than leading with good news, which answers a question nobody asked.

What is usually missing from these messages?

The list of what is not transferred. Teams describe what the migration covers and go quiet about the gap, because it is awkward. Name it: local files outside synced folders, anything installed without administrative rights, stored credentials, virtual machines, container images. One sentence with a deadline prevents the conversation that starts with nobody told me.

Should we announce it company-wide or individually?

Both, in that order and separately. One company note explaining that a refresh cycle exists, what the interval is, and how people will be told. Then an individual message with each person’s own date. A company note with no personal date creates anxiety, and a personal note with no context reads as if something is wrong with their machine in particular.

How long should people keep both machines?

Two days for most people and about four for anybody with a heavy local setup. A same-day handover means they stop work, migrate, find something missing and cannot retrieve it. Overlap turns each of those discoveries into a two minute fix, and the cost is holding one extra device per person, staggered across the programme.

What should happen at the end of the overlap?

Ask the person to confirm they have what they need, then book collection immediately. That one line converts the return from something being taken into something being released, and it removes the commonest reason machines sit uncollected for months, which is that the person was not sure they had finished with it so they quietly kept it.

What if somebody refuses the new machine?

Find out why first, because the reasons are usually practical. Bad timing mid-project is reasonable and solvable by moving the date. An irreproducible dependency is a real technical risk needing a technical conversation. A machine that is genuinely worse for their work is the same signal as an upgrade request. Hold the line only on devices past end of support, and call that retirement rather than upgrade.