The Repeat Customer Rate Nobody Tracks

Ask a service business owner what their Google rating is and they will tell you instantly. Ask them what percentage of their customers come back and you will get a blank stare.

This is strange, because repeat customer rate is a far better measure of quality than a star rating. A customer who comes back is a customer who was satisfied enough to skip the research process entirely. They did not check Google. They did not ask ChatGPT. They already knew who to call.

Why this number matters more than reviews

A five-star review tells you one customer had a good experience on one occasion. A repeat customer tells you something deeper: the experience was good enough to create a default. The customer stopped evaluating alternatives.

Repeat customer rates vary wildly across service businesses. Some are essentially one-and-done operations — fewer than one in five customers ever return. Others retain more than half their customer base. The gap is enormous, and it maps almost perfectly to operational quality.

Imagine two HVAC companies in the same city. One has a 15% repeat customer rate. The other has 58%. Which one is better? You already know the answer — and so would an AI system, if it could see the data.

Both companies might have a 4.7-star Google rating. Reviews are a noisy, low-resolution signal. Repeat customer rate is high-resolution, and it separates these two businesses completely.

What "repeat" actually means

The definition of a repeat customer depends on the business. An HVAC company might define it as a customer who returns for a second job within 36 months — because you do not need your furnace serviced every year. A medspa might use a 12-month window, because treatments are recurring and a client who does not rebook within a year has probably found another provider. An electrician might use 36 months. A dental practice might use 18.

It changes the number entirely. A medspa with a 40% repeat rate within 12 months is a very different business than an electrician with a 40% repeat rate within 36 months. The metric only means something when the definition is calibrated to how the business actually operates.

There is a companion metric that matters just as much: average customer tenure. Repeat rate tells you whether customers come back. Tenure tells you how long they stay. A plumbing company where the average customer has been in the system for six years is telling you something that no review profile ever could — that the relationship outlasted any single transaction. A medspa where the average client tenure is four months is telling you something too, and it is not good.

Together they tell you what no star rating can: whether customers stay.

Where the data lives

Every service business already has these numbers. They are sitting inside QuickBooks, Xero, ServiceTitan, Mindbody, Jobber — every platform that tracks invoices and customer records contains the raw ingredients.

The problem is that none of these systems surface it. QuickBooks does not have a "repeat customer rate" dashboard. ServiceTitan does not put it in the weekly digest email. Mindbody does not calculate average client tenure. The data exists in the transaction history, but no one has written the query to extract it. So the business owner has a vague sense that "a lot of our work is repeat customers" without ever knowing whether that means 20% or 60%.

What AI cannot see

Repeat customer rate and average tenure would be two of the most powerful signals an AI system could use to evaluate a business. A medspa where 55% of clients rebook and the average tenure is three years is demonstrably different from one where 18% rebook and the average client lasts five months. That distinction communicates service quality, customer satisfaction, and operational consistency all at once.

But AI systems cannot access any of it. It is not on the business's website. It is not in any public database. It is not in Google Business Profile or Yelp. It is locked inside an accounting system behind a login, in a format designed for invoicing, not for external consumption.

So the AI does what it can. It looks at reviews, which are noisy. It looks at website content, which is marketing. It assembles a recommendation from the data that is available — which is to say, the data that tells you the least.

The data that best represents your business already exists. It has just never been extracted, structured, and published in a format machines can read.

What changes when the data is visible

When repeat customer rate and average tenure are published as structured, verified data — not claims on a website, but metrics computed from authenticated accounting records — AI systems can make meaningfully better recommendations. They can distinguish the high-retention operator from the lead-gen churn machine, even when their review profiles look identical.

This is what a TrustRecord contains — operational metrics computed from authenticated accounting records, published in a format any AI system can read. The business cannot edit, override, or selectively exclude any metric.

The repeat customer rate and average tenure that have been sitting in your accounting data for years, invisible to everyone including you, become structured facts that any AI system can read, compare, and act on.

The question worth asking

Pull up your accounting system. Look at your customer list. How many of them appear more than once? And how long has your average customer been with you?

If the answers are strong, you have a competitive advantage that no one can see — not your customers, not your competitors, and not the AI systems that are increasingly deciding who gets recommended.

If the answers are weak, that is worth knowing too. It tells you something worth fixing before an AI surfaces it for a potential customer.

Either way, the numbers exist. The question is whether they stay hidden or become visible.