Six steps of retention: find yours before you buy anything

Ask ten founders what retention means and you get ten answers. For one it is four welcome emails. For another it is a system that predicts who is about to leave and picks a different offer for each person. Same word, very different money. Here is the six-step ladder I use to tell which one you are talking about.

What are the six steps?

Step 0: nothing. Customers leave and nobody notices until revenue drops.

Step 1: a sequence. Welcome emails, a birthday discount. The same message for everyone, sent on a timer.

Step 2: triggers. A message goes out because something happened: a missed visit, a failed card, a cart left behind. New customers and regulars hear different words.

Step 3: several channels, one brain. Email, SMS, WhatsApp, a phone call. The channel depends on what happened and who the person is, and there is a limit on how often one person hears from you.

Step 4: prediction. A model estimates who is likely to leave and what they are worth, so the strongest offer goes to the people where it matters.

Step 5: proof in money. A small group gets nothing. You compare. Now you know what the system earned after paying for the discounts.

Where does a typical small business sit?

Most businesses I talk to sit on step 1 or 2. Some are on step 0 and do not know it, because the newsletter goes out every month and feels like activity.

They also want to jump to step 4, because it sounds like AI. I understand the pull. A tool that says it will predict who leaves is easier to buy than a habit of writing down due dates. But a prediction needs history, clean data and someone to act on the list, and without those it produces confident nonsense. I say more about that in what machine learning does in retention.

Why is step 5 the one to do first?

Because it is the cheapest step on the list and the only one that tells you whether the others earn money.

Without it, you can build triggers, add channels and buy a prediction tool, and still not know whether any of it changed what customers did. The reports will show messages sent, opens, clicks and orders with attribution. Every one of those counts people who would have come back anyway.

With it, each step above becomes testable. You add a trigger, hold back one customer in ten, and see whether the trigger moved anything. If it did not, you saved yourself the cost of the next three steps.

The method is in how to know a campaign worked. It takes one decision before the next send, and it needs no software.

How do you find your step in two minutes?

Answer five questions, yes or no, in order. Stop at the first no.

1. Do you know, for each customer, roughly when they should be back? If not, you are on step 0.
2. Does a message go out automatically on a schedule or at a fixed moment after purchase? If yes, you are at least on step 1.
3. Does a message go out because of something the customer did or did not do, such as a missed visit? If yes, step 2.
4. Do you choose the channel and timing by the customer, within a limit on how often you contact anyone? If yes, step 3.
5. Does a model rank customers by who is likely to leave or return? If yes, step 4. And do you compare against a group you did not contact? If yes, you are on step 5.

Be strict. If you send triggers but never compared with a group, you are on step 2, whatever the tool says.

Does everyone need to climb all six?

No. A single-location barber with 300 regulars may be well served by step 2 and a proof at step 5. A larger business with tens of thousands of customers and several channels gets more from steps 3 and 4. A model costs money to build, to retrain and to watch, so it pays when the customer base is large enough to feed it and the value per customer is high enough to repay the work.

The ladder asks you to know where you are, and to check that your current step earns its keep before you pay for the next one.

Should I buy software or build a spreadsheet?

It depends on how many customers you have and how much time the counting takes. I wrote a comparison for subscription businesses, DIY spreadsheet, one-time report or retention software, that lays out when each makes sense. The same logic applies elsewhere: start by hand until the hand work is the bottleneck, then buy.

Why do the steps matter for budget?

Because each step costs more than the one before it, and the cheapest step with proof often beats an expensive one without. Say you spend $500 a month on a tool that sends triggers and you have never held anyone back. You do not know whether that $500 earns $2,000 or nothing. An afternoon of setup on step 5 would tell you. This is an example, but the logic holds for any amount.

What can you do today?

Run the five questions above and write down your step. Then choose one of two moves.

If you are on step 0 or 1, add a "next due" column and send three personal messages this week to the customers most overdue.

If you are on step 2 or higher, pick the campaign that runs most often and hold back a random tenth of customers from the next send. Compare in a month. That is step 5, and it costs you nothing but a few sales you may not have lost.