Automating a Follow-Up Without Being Creepy

The line is real

A follow-up after a quote is normal business and welcome. A message arriving eleven minutes after somebody looked at a page, naming the page, is not.

Both are automated, both are technically the same mechanism, and they land completely differently.

What separates them is speed, precision, and whether the message is useful on its own.

Which means the discomfort is avoidable, and it is created by settings rather than by the idea of automation.

What makes it uncomfortable

• Speed: arriving within minutes.
• Precision: naming exactly what they did.
• Asymmetry: you know something they did not share.
• Repetition: a third message about the same thing.
• Pretending: writing as though it were personal.

The fifth is the one businesses get wrong while trying to be warm. A templated message written as though somebody sat down to type it is worse than an obviously automatic one.

Delay is most of the fix

The single change that resolves most of the discomfort.

A message arriving minutes after an action makes the watching obvious. The same message the next morning does not.

For a quote follow-up, two or three days is normal and reads as diligence.

For anything triggered by browsing rather than by an action somebody took deliberately, at least a day, and preferably longer.

The delay costs almost nothing in response and removes almost all of the objection, which is an unusually good trade.

Where a tool defaults to sending immediately, that default is set for a different kind of business.

Be general about what you know

The second change, and it is a wording decision rather than a technical one.

You saw our bathroom renovation page on Tuesday afternoon is precise and unsettling.

You were looking at bathroom work, so here is what people usually ask us next is the same trigger, generalised, and it reads as relevance.

The information used is identical. What changes is whether the message demonstrates surveillance or simply appears well timed.

As a rule, reference the topic rather than the action, and never reference the timing.

Never mention that they opened a previous email, which is the version that most reliably produces a complaint.

Make it useful on its own

The test that separates a follow-up from a nudge.

If the only content is a reminder that they did not finish something, the message exists for you rather than for them.

Add the thing they would have wanted anyway: what the job usually costs, how long it takes, what to check before deciding, what people ask most often.

That way somebody who has decided against you still gets something, and somebody who was hesitating gets the information that resolves it.

It also means the message can be sent honestly to people who never respond, since it was worth sending regardless.

A worked example

A business set up a follow-up for anybody who requested a quote and did not respond within a week.

The first version arrived after two days, opened by naming the exact service and the date of their enquiry, and asked whether they had made a decision.

Two people replied to ask how the business knew they had not decided, which was a reasonable question about a message that had implied more monitoring than existed.

They rewrote it: sent after four days, from the owner's address, opening with the two questions people usually ask about that kind of job, and closing with an offer to answer anything.

Response roughly doubled and the uncomfortable replies stopped.

The trigger, the list, and the timing window were nearly identical. The wording had been the whole problem.

Say that it is automatic

A counterintuitive point that consistently helps.

A short line saying this is an automatic message, and inviting a reply if it is not relevant, removes the pretence and with it most of the discomfort.

People are entirely comfortable with automation they can see, and uncomfortable with automation disguised as attention.

It also protects you from the awkward case where somebody has already spoken to you and receives the message anyway, which happens in every automated sequence eventually.

One sentence, near the end, and it costs nothing.

Stop when they respond

The mechanical requirement that most sequences get wrong.

Somebody who replies, telephones, or books should be removed from the sequence immediately, and in many setups they are not.

A customer who has already booked receiving a message asking whether they are still interested is the version of this that damages a relationship.

Check that your tool removes people on reply, on booking, and on purchase, and test it by going through the sequence yourself.

Where the systems do not connect, a shorter sequence is safer than a longer one, since each additional message is another chance to contact somebody who has already dealt with you.

Two messages is usually the limit

On length, since sequences tend to grow.

For a small business, a follow-up and one further message is enough, and the second should be lighter than the first.

A third message about the same enquiry is where persistence turns into pressure, and the return on it is small.

Where somebody has not responded to two useful messages, they have answered, and the remaining route is a phone call from a person or nothing.

Ending a sequence with a plain message saying you will stop contacting them about this is a better last impression than trailing off.

The counter-case

Automated follow-up is genuinely valuable and this should not read as discouragement.

Most quotes that go unanswered go unanswered because the customer got busy, not because they decided against you, and a follow-up recovers a meaningful proportion of them.

Doing it manually is the alternative, and manual follow-up is the thing every small business intends and few maintain.

There are also businesses where speed is the service, and a fast response to a specific action is exactly what somebody wants rather than a surprise.

Automate it, delay it, keep it general, make it useful, and make sure it stops.

Setting it up

1. Delay it by days rather than minutes.
2. Reference the topic, never the action or the timing.
3. Never mention that they opened something.
4. Include something useful beyond the reminder.
5. Say it is automatic in one line.
6. Test that it stops when somebody responds.
7. Keep it to two messages.

Step six is the one worth testing rather than assuming, since a sequence that keeps running after somebody has booked is the version that costs you a customer.

Which messages are worth automating is covered in automatic emails worth setting up once.

Frequently asked questions

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.