How do you calculate customer retention rate?

Customer retention rate = [(customers at end of period − new customers) ÷ customers at start of period] × 100. The formula fits on one line; what usually goes wrong is the choice of period and the way you read the result. This guide walks through the calculation with a worked example, benchmarks by product type, and the actions that actually keep your customers.

What is customer retention rate?

Retention rate measures the share of your existing customers who are still customers at the end of a given period. It is the mirror of your dependence on acquisition: the lower it gets, the more your revenue rests on customers you have to pay to replace.

The financial stakes have been documented for a long time: according to research by Bain & Company, increasing retention by 5 points can grow profits by 25 to 95%. The mechanics are simple: a retained customer carries no acquisition cost, orders more often, and orders bigger over time. When ad costs climb, retention becomes the cheapest growth engine you can pull.

How do you calculate your retention rate?

The formula: retention rate = [(customers at end of period − new customers from the period) ÷ customers at start of period] × 100. You subtract the new customers from the final total so you only measure the loyalty of existing customers, without recent acquisitions artificially inflating the result.

The formula with a worked example

Your store counts 500 active customers on January 1. On March 31, you count 480, including 50 acquired during the quarter. Calculation: (480 − 50) ÷ 500 × 100 = 86%. Out of 500 starting customers, 430 stayed, 70 left.

Churn, or attrition rate, is the exact inverse: 100 − 86 = 14% of customers lost over the period. Both numbers tell the same story, so pick the one your team tracks most naturally. On the tooling side, Shopify shows the "returning customer rate" in its customer reports, and an order export with emails is enough to rebuild the calculation in a spreadsheet.

Pick the right period: your repurchase cycle

A measurement period disconnected from the buying rhythm produces meaningless numbers. A coffee subscription seller can measure retention monthly: their customers order every four weeks. A mattress seller doing the same would conclude that 99% of their customers have "left", when they simply have no reason to buy again for years.

The rule: your measurement period covers at least one full repurchase cycle. Consumables and beauty: the quarter. Fashion: the half-year. Durable goods: the year, and the referral rate often becomes more telling than retention itself.

Cohort retention, the reading that tells the truth

The overall rate smooths everything out. The cohort reading, on the other hand, shows where the relationship breaks: take the customers acquired in a given month, and track what share bought again at 30, 60, 90 days. If the January cohort retains better than the March one, something changed between the two (a product line, a carrier, a wave of less qualified traffic). Most stores discover this way that their retention plays out in the first 60 days: a customer who has not bought again by then will probably never buy again. That is where the actions should concentrate, not on customers who are already settled in.

Retention rate, loyalty, churn: what are the differences?

The three terms overlap without being the same thing. Retention rate is a measurement: the share of customers kept over a period. Churn is its complement: the share of customers lost. Loyalty is the work that produces those numbers: loyalty program, service quality, the relationship after the purchase.

Put differently, retention is the thermometer, loyalty is the treatment. Tracking the first without investing in the second amounts to watching the temperature drop without changing anything. The opposite mistake exists too: multiplying loyalty initiatives without measuring retention, and therefore without knowing what works.

What is a good retention rate in e-commerce?

There is no universal good rate: everything depends on your category's repurchase cycle. The ranges below give orders of magnitude in annual retention, compiled from public benchmarks published by e-commerce platforms.

If rebuilding cohorts feels heavy, start with a simpler indicator: the share of orders placed by existing customers (Shopify's "returning customer rate"). It does not replace the retention rate, but it reads in one click and moves in the same direction.

The most useful comparison remains your own history: a retention rate that climbs quarter after quarter beats a flattering benchmark. And always cross it with your average order value: fewer customers who spend more per order can mask an erosion of the base.

To complete the picture, here is how to calculate your conversion rate: the formula, the right denominator and the segmentation that makes the number readable.

How do you improve your retention rate?

Customers almost never leave on a whim. They leave because nothing holds them, or because a bad experience went unanswered. The two are worked on differently.

The basics: post-purchase that keeps the relationship going

The period between two orders decides the next one. A useful post-purchase email (usage tips, not a promotion), clean delivery tracking and a simple loyalty program are enough to maintain the link. The loyalty program plays a double role: accumulated points give a reason to come back, and earned status a reason not to leave. We detailed this mechanism in our guide on customer loyalty through reviews, repurchase numbers included.

Listen to unhappy customers before they leave

An unhappy customer who says nothing is already almost lost: they do not complain, they do not come back. Asking for a review systematically after every order turns that silence into a signal. The dissatisfied customer who receives a review request has somewhere to speak up, and you have a chance to win them back before they disappear. That is the principle of preventive mediation: detecting dissatisfaction the moment it appears, not six months later in the statistics. According to an Ifop study from January 2026, 58% of consumers prefer businesses that respond to reviews: listening shows, and it retains.

Measure what works: retention by review segment

To know whether your listening setup actually retains, segment your retention into three groups: customers who left a positive review, those who expressed a dissatisfaction your team handled, and those who never responded. The third group serves as the control. At most merchants, the gap is clear: a customer who was contacted back after a problem buys again more often than a silent customer, because the relationship existed. If your "handled complaints" segment does not retain better than the control, the problem is not the collection, it is the handling behind it.

Solution

What if your unhappy customers gave you a second chance?

Review Collect alerts your support team the moment a customer rates 1, 2 or 3 stars. 93% of unhappy customers contacted prefer to resolve privately.

Discover Review CollectBook a demo

×30 reviews collected in the first month

How Review Collect improves your retention

Review Collect turns review collection into a dissatisfaction radar. Every order triggers a review request by SMS or WhatsApp, with a 40% response rate on average: you finally hear the customers who would never have told you anything. When a customer rates 1, 2 or 3 stars, they land on a feedback page in your brand's colors, and your support team receives an alert with the text, the order number and the contact details. Your team calls back, the problem gets solved.

93% of customers contacted after a bad experience prefer to resolve privately rather than post a review.

A customer won back at that moment is not just a negative review avoided: it is a customer who saw that their problem mattered, and who has a real reason to recommend you. Review collection is up and running in 48 hours, without a developer. Your retention rate is in your back office; the customers on their way out are in your next collection campaign.