The loss happens after the lead arrives
A marketing team gets judged on volume. Enquiries this month against enquiries last month, cost per enquiry, which channel produced them. All of that is counted carefully. What happens to an enquiry in the hours after it lands is counted by nobody, and that is where the spend goes back out of the building.
The pattern repeats across the teams we have sat with. Enquiries arrive by web form, by email, by phone and through social messages, and they arrive at every hour of the week. Someone copies them into a spreadsheet. The reply goes out at the next inbox check, which on a Friday evening means Monday. A quote goes out, one chase follows a few days later if the person remembers, and the rest of the open quotes sit there until they expire.
The teams doing this are not careless. They are running a process that depends on a person being at a desk and remembering, applied to something that arrives at random hours and rewards whoever answers first.
Why this one goes first
It works demand the company has already paid for. Before a budget goes towards more enquiries, the ones that arrived and went unanswered are the cheapest revenue in the business, and nobody has to approve a new channel to go after them.
It is also small. The tools are already paid for in most companies: a CRM, a form, an email sender, a scheduler, and often an SMS account that came with the phone system. The build is mostly configuration and writing, and the writing takes longer than the configuration.
And it shows. A team can watch the first automated reply go out on the evening of the day it went live, and at day thirty the reply time to three test enquiries is a figure they can read for themselves. Reporting, CRM hygiene and the operations work that tends to follow are all worth more across a year, and none of them prove themselves that quickly. Follow-up buys the credit that the rest of the work runs on.
What we build
Every enquiry becomes a record when it arrives, whatever it arrived through: a web form, an email to the shared inbox, a missed call, a message on a social account. The record carries the source and the time it came in, which makes the campaign reporting honest later on as a side effect.
An acknowledgement goes out at once, from a named person, with a booking link in it. A few qualifying questions sit either in the form or in that first message, so need, budget and timing are on the record before anyone spends a call establishing them.
The sequence itself is deliberately short. Any enquiry the contact has not answered after an agreed number of days gets a small run of messages, and so does any quote still sitting open. Both stop the moment the contact replies or books, and a reply goes to a person and not back into the automation. Deals move stage on their own as a quote is viewed, accepted or paid, so the pipeline stays accurate without anyone updating it on a Friday.
None of that is technically hard.
The wording takes longer than the build
Configuration is the short part of the job. The arguments are about the words, and they are worth having. A first reply written like a receipt gets ignored. A first reply promising a call within the hour commits somebody to answering the phone at nine on a Sunday evening, which is a decision for the person whose phone it is.
So the messages get written with the person whose name is on them, in the way that person actually writes, and the number of them is kept low enough that they would be comfortable sending every one by hand. That's the test. If the person named at the bottom would not send it themselves, it doesn't go in the sequence.
The same applies to what the automation says when it has nothing useful to add. A third message that exists only because the sequence had a slot for it is the message that gets a company marked as spam, and one of those costs more than the enquiry it was chasing.
The rules that keep it from becoming noise
Automated follow-up earns its bad reputation when the rules for stopping are weaker than the rules for sending. The ones we hold to are ordinary. The sequence stops on any reply, on a booking, and on an unsubscribe. The number of messages is capped, the cap is agreed with the client, and it is low. Every message comes from a named person with a working reply address, and every message carries one ask.
The consent basis and the date go on the record, because a follow-up programme that cannot say why it is writing to someone is a complaint waiting to happen. Bounces and unsubscribes are dropped by a rule inside the system, so nobody has to remember.
The whole thing is written so that a person in the client's team can change it. When the offer changes and the sequence keeps describing the old one, the automation does damage at a steady rate and nobody is watching. That is why the handover names the trigger and then every step in the tools it runs in.
