yodai Academy Analytics

Academy › Publishing › Analytics

What counts as "they came back" in your business

You'll walk away with your own rule, written in one sentence, for what counts as a customer coming back, and what doesn't count even if it looks like it.

By Yodai · Published September 6, 2026 · Updated October 12, 2026

7 min read

You look at the month's calendar and see Marta listed twice. The first was a haircut. The second was actually the same appointment moved because of rain. And Carmen bought a package of five physical therapy sessions and has already used the second one: does that count as Carmen "coming back," if she already paid for all five at once on day one?

Before you can calculate how many of your customers are repeat customers, you need to settle something much simpler, and easy to skip: what counts as "coming back" in your specific business. The answer isn't the same for a hair salon, an auto shop, and a physical therapy clinic.

If you don't set that rule yourself, any number you come up with later will be counting different things without you realizing it. And a number that mixes different things together isn't useful for deciding anything.

What doesn't count, even if it looks like it does

There are three things that sneak onto the list of "repeat customers" without actually being repeat customers.

The first is a rescheduled appointment. If someone switches Thursday for Friday, it's still the same appointment and the same payment: they didn't decide to come twice, they decided to come once on a different day. If your calendar counts every reschedule as a new visit, you'll see repeat business where there's really just a time change.

The second is a session from a package you already got paid for. If you sold five physical therapy sessions upfront, each of those five visits is the same customer using up what they already paid you on day one, not a new decision to come back to you. Logging all five as "five times they came back" inflates the number without anything new actually happening in between.

The third is the same customer entered twice under two different spellings of their name. Mari and María José Fernández could be the same person with the same phone number. If your criteria are based on how the name is written instead of the phone number or email, you'll count people who already knew you as new customers, and your repeat rate will come out lower than it really is.

Counting package sessions as if they were repeat business
Here's a costly mistake that's easy to make: logging every visit from an already-sold package as a new repeat sale. That way, a business that actually sold only ten packages of five sessions each can look like it has fifty repeat visits. The number goes up, but it doesn't tell you whether you're winning loyal customers or just delivering what you already got paid for weeks ago. If you sell packages, count each package as a single sale for repeat-customer purposes, never as five separate sessions.

The question that sets your criteria

To figure out whether a payment counts as the customer coming back, ask yourself just one question: is this a new decision to come in, or is it the continuation of something they'd already decided before?

A new decision is when someone, with nothing owed to them and nothing left pending, chooses to pay you again. That's coming back. Using up a package, picking up an order they already paid for, or rescheduling an appointment doesn't count, even though it looks like just another visit on the calendar.

Once you've settled that question, there's a second decision left, and this one doesn't have a single right answer: what counts as a payment to you. A customer who comes back for exactly the same thing isn't the same as one who comes back for something different, as long as they're still buying it from you.

Any payment to you, or only the same service
Count any payment to you
  • A customer who came in for a haircut and comes back for a color treatment counts the same as one who repeats the same haircut
  • This works when your services are all connected by the same trust: if they come back, it's because they trust you, not because of the exact service
  • It's simpler to track: just one column, "came back: yes or no"
Count only the same service
  • Only buying exactly the same thing counts as a repeat: another oil filter, not a set of new tires
  • Requires tracking a separate column for each type of service or product, not just one

Which one do I pick? If what you sell revolves around a single relationship built on trust, your hair salon, your clinic, your auto shop, count any payment from that person as a return visit: what's repeating is trust in you, not a specific service. If you sell unrelated categories, count by category: this works when each purchase is its own separate decision, since buying a frying pan says nothing about whether that person will come back for a phone charger.

How to set your criteria this week

How to set your criteria this week
  1. 1Review your last twenty appointments or sales
    You don't need to look at the whole year. Your last twenty will already show you the odd cases: reschedules, half-used packages, duplicate names with typos.
  2. 2Identify each customer by phone number or email, not by name
    If two rows share the same phone number, they're the same person, even if the name is spelled two different ways.
  3. 3Decide whether any payment to you counts or only the same service
    Use the question from before: if your services share the same trust, count any of them; if they're unrelated, count by category.
  4. 4Write the rule down in one sentence and don't touch it every month
    Once you've set it, apply it the same way every time. If you change it partway through the year, the numbers from before and after stop being comparable.
The written rule, for a hair salon
A customer counts as having come back when: they pay me a second time, whatever the service, a cut, color, or manicure, identified by their phone number, as long as it's not a reschedule of the same appointment or a session from a package I already got paid for in full.

With that sentence written down, you have what you need for the next step: counting, out of every ten customers, how many meet that rule and actually come back.

What you'll walk away with
  • You have your own rule, written in one sentence, for what counts as a customer coming back
  • You know what doesn't count even though it looks like it does: reschedules, sessions from a package already paid for, and the same customer with their name spelled differently
  • You identify each customer by phone number or email, not by however their name happens to be spelled
  • You've decided whether to count any payment to you or only the same service, and why

Test yourself

Answer before opening each one. If one gives you trouble, go back to the section that explains it.

1. A customer at your repair shop brings their car in Monday for a checkup, and comes back Thursday because you forgot to change a filter and they have to stop by again, at no cost.

This doesn't count as them coming back. There's no new decision to pay you: it's the same checkup that wasn't fully wrapped up, even though the calendar shows two appointments on different days.

2. A customer bought a package of ten physical therapy sessions, used them all up, and two months later comes back and buys another package of ten.

This does count. Buying the second package is a new decision to pay you again, different from simply working through the sessions in the first one, which were already paid for from the start.

3. Your hardware store sold a customer a drill in March. In July, that same customer comes back and buys paint for their walls.

It depends on the rule you set, not the product. If you count any payment from them because what your hardware store really sells is trust in the place, that counts as a returning customer. If you count by category because each purchase is its own decision, it doesn't count as a repeat "tools" purchase, even though it's still a new sale. What matters is applying the same rule every time, not changing it from month to month.

With that rule, you can already filter your customer list. What you still don't know is what percentage of them, overall, come back: you'll calculate that in the next lesson, along with the mistake that calculation carries if you don't watch which period it's drawn from.

NextOut of every ten, how many come back