Aftercare instructions: sent at the moment they are wanted

Aftercare is handed over at the worst possible moment. Somebody has just had a treatment, is putting a coat on, is thinking about parking, and receives a leaflet plus a verbal summary of what to do for the next fortnight. Almost none of it lands, and the leaflet is in a coat pocket by the evening.

The information was fine. The timing was wrong. Most of what aftercare tells people is relevant on day three, day seven and day fourteen, and delivering all of it at hour zero is why the same questions come back as phone calls later.

What we'll cover

• Why the handover moment fails
• Staging it to when each part matters
• The first message, sent before they leave
• The check-in that earns its place
• Building it with CodeWords
• Writing instructions people follow
• Stopping it when it should stop
• More than one sequence
• Where the review request belongs
• What has to reach a person
• Making it survive
• Limits worth knowing about
• What to build first
• Frequently asked questions

Why the handover moment fails

Three things are working against it and none is about the quality of the advice.

Attention is elsewhere. Somebody who has just been treated is processing the experience, not encoding a schedule.

Most of it is not yet relevant. Instructions about week two cannot be acted on in week one, so they are set aside and then forgotten.

Paper leaves. It goes into a pocket, a bag, a bin. The phone does not.

Which means the intervention is not a better leaflet. It is sending each piece at the point the person needs to act on it, in a place they still have.

Staging it to when each part matters

Take your existing aftercare sheet and sort every instruction by when it applies. Most fall into four groups.

Immediately and today. What to avoid in the first hours, what discomfort is expected, what to do tonight.

The next few days. The bulk of it. What to apply, what to avoid, what normal recovery looks like.

The point things change. When a dressing comes off, when normal activity resumes, when a product can be used again. This is the instruction most commonly missed, because at hour zero it was too far away to register.

The end. What the finished result should look like, and what to do if it does not.

Each group becomes one message at its own time. Nobody reads a long document; everybody reads a short message that is about today.

The first message, sent before they leave

One habit improves everything downstream: send the first message while the person is still with you and ask them to check it arrived.

It confirms you have the right number, which is otherwise discovered three days later when nothing has been delivered. It establishes the channel as the place aftercare lives. And it means they leave with the information on a device rather than in a hand.

Keep it short and make it about right now. The rest of the sequence can carry the detail.

The check-in that earns its place

At the point where problems typically appear, ask a direct question rather than sending more instructions.

The value is in the reply. Somebody who says it is uncomfortable, or that it does not look right, has told you something you would otherwise hear as a complaint in a fortnight or not at all. Catching that early is better for them and considerably cheaper for you.

Two conditions make it work. The question has to be specific enough to answer in a word or two, and the answer has to reach a person quickly. A check-in nobody reads is worse than none, because it invites somebody to raise a concern into a void.

Place it where your own experience says problems surface. That timing is specific to what you do, and your records already contain it.

Building it with CodeWords

CodeWords connects to more than 3,000 integrations, and the connection is made once and reused.

1. Open CodeWords and start a new automation.
2. Describe what should happen in plain language to Cody, the automation builder: which treatment triggers which sequence, the timing of each message, and which replies must reach a person.
3. Authorize the connections to your booking or records system and your WhatsApp Business Platform account.
4. Describe the exceptions: a cancelled appointment, a reply expressing concern, somebody asking to stop, a client who has already been in touch.
5. Run it on one treatment type for a month before extending it, and read every reply yourself while you do.

You describe the outcome; Cody builds it, connects it, and deploys it. The free plan covers light use, with Pro at $39 per month and Business at $100 per month as usage grows; details are on the pricing page.

Writing instructions people follow

Most aftercare text was written to be defensible rather than to be followed.

One instruction per line. A paragraph containing four things gets treated as one thing and half of it is lost.

Say what to do, not what to avoid, wherever both are possible. "Keep it dry for 24 hours" beats "do not allow it to get wet".

Name the time. Tonight, tomorrow morning, Thursday. Relative periods get miscounted.

Say what normal looks like. Most anxious calls are somebody wondering whether what they are seeing is expected, and a sentence describing normal prevents a large share of them.

Put the concerning signs in their own message, not buried in a list of routine advice.

Stopping it when it should stop

A sequence that keeps sending after the situation has changed is the fastest way to lose the channel.

Stop on a cancellation or a no-show. Aftercare for a treatment that did not happen is a small disaster.

Stop on a reply, and route it. Somebody in a conversation should not also be receiving scheduled messages.

Stop when they ask. Immediately, permanently, and across every automation rather than just this one.

Do not restart it for a follow-up appointment unless that appointment genuinely has its own aftercare.

More than one sequence

One sequence for everything is the usual starting point and it stops being right quickly.

By treatment. Different work has genuinely different instructions and genuinely different timings. Sending one generic sequence means most people get advice that does not apply to them, and they learn to skim.

By risk. Somebody you have flagged for any reason, or who had a longer or more involved appointment, warrants an earlier check-in and more explicit instructions about what to watch for. Your records already know which appointments those were.

By whether they have been before. A returning client does not need the orientation a first-timer does, and sending it anyway makes the whole thing feel automated in the way people dislike.

Start with one sequence for your most common treatment, then split it when you can point at what would actually differ. Splitting speculatively produces several sequences nobody maintains, which is worse than one that is slightly too general.

Where the review request belongs

It does not belong inside the aftercare sequence, and putting it there is common.

Aftercare is a duty of care. A review request is a favour you are asking. Mixing them makes the care look like a pretext, and people notice.

Send it after the sequence has finished and the outcome is known. Never send it to somebody who raised a concern, until a person has resolved it. And check the rules of the platform you are asking about, since several prohibit selectively soliciting only from people who seemed pleased.

What has to reach a person

Anything describing a symptom. Pain, swelling, bleeding, a reaction. An automated reassurance here is unsafe and reads badly afterwards.

Anything that sounds like a complaint. Route it immediately, with no automated acknowledgement that could seem dismissive.

Any question the sequence did not anticipate. Better a slower human answer than a confident wrong one.

Silence from somebody who said something was wrong. A concern raised and not followed up is worse than a concern never invited.

Making it survive

Check the appointment still happened before every send.

Keep the content with whoever is responsible for it, and date it when it changes.

Route replies to a person who is actually available, with a named owner.

Count the replies by type, since they tell you which instructions are unclear.

Review the timing annually against when problems actually appeared.

Limits worth knowing about

A message cannot assess anything. It delivers instructions and prompts; judgement is a person's.

Some people will not read any of it, and the phone call is still part of the job.

More messages are not better care. A sequence that feels like marketing gets muted, and then the important one is muted too.

Timing is trade-specific. Copying another business's schedule gets it wrong in both directions.

The channel has rules. Outside the 24-hour window after somebody messages you, contact requires an approved template, so the sequence has to be built with that in mind.

What to build first

The first message, sent before they leave, confirming the channel and covering today only. One message, one treatment type. It fixes the wrong-number problem, it establishes where aftercare lives, and it is the foundation everything else sits on.

Two habits make the difference. Sort your existing instructions by when they apply rather than by topic, since that single reordering is most of the value here. And make sure the check-in reply reaches somebody who will read it, because inviting a concern you do not answer is worse than never asking.

Frequently asked questions

Why do people forget aftercare instructions?

They receive them at the moment they are least able to absorb them, and most of the content is not relevant until days later. Paper also leaves with them and is then lost.

How should an aftercare sequence be staged?

Sort every instruction by when it applies: today, the next few days, the point something changes, and the end. Each group becomes one short message sent at its own time.

When should the first message go?

Before the person leaves, with a request to confirm it arrived. That verifies the number and establishes the channel while they are still in front of you.

Is a check-in message worth sending?

Yes, if somebody reads the replies. Timed to when problems typically appear, it catches issues early. Inviting a concern nobody answers is worse than not asking.

What should aftercare instructions look like?

One instruction per line, phrased as what to do rather than what to avoid, with named times rather than relative ones, and a clear description of what normal looks like.

When should the sequence stop?

On a cancellation or no-show, on any reply, and immediately and permanently when somebody asks, across every automation rather than just this one.

Should a review request be part of aftercare?

No. Aftercare is a duty of care and a review request is a favour. Send it afterwards, never to somebody who raised a concern, and check the review platform's rules on selective solicitation.

What must reach a person?

Anything describing a symptom, anything resembling a complaint, any unanticipated question, and any silence following somebody saying something was wrong.

How many messages is too many?

Enough that it starts to feel like marketing. A muted sequence takes the important message with it, so fewer and better-timed beats more. Four or five well-placed messages generally does more than a daily one.

Should there be different sequences for different treatments?

Yes, once you can say what would actually differ. A single generic sequence means most people receive advice that does not apply to them and learn to skim, which costs you the messages that mattered.

Can I copy another business's timing?

The timing should come from when problems actually appear in your own work, which your records will show. Borrowed schedules are wrong in both directions.

Related reading

• The clinical boundary on recovery advice
• Designing the conversation
• The edge cases that actually matter