Customer Retention Rate: Formula and How to Improve It

Customer retention rate is the percentage of customers you keep over a period, and it determines how long customers stay, how much they are worth and how much you must spend on acquisition just to stay the same size. The formula is simple, but the inputs need care: you must exclude customers acquired during the period, define "retained" in a way that fits how your customers actually buy, and convert between monthly and annual figures correctly. This guide gives the formula, worked examples, a section on retention for non-subscription businesses, the effect of retention on lifetime value, and a list of ways to improve it. It belongs to unit economics and budgeting and forecasting, builds on cohort analysis, and leads to churn rate and revenue churn vs. logo churn.

The formula

Retention rate = (Customers at end of period - New customers acquired in period) / Customers at start of period

New customers are excluded because the question is how many of the customers you started with are still there.

Example: a company starts a quarter with 2,000 customers, acquires 500 new ones, and ends with 2,300.

• Customers retained = 2,300 - 500 = 1,800
• Retention rate = 1,800 / 2,000 = 90%
• Customers lost = 2,000 - 1,800 = 200, so churn is 10%

A common error is to divide the ending count by the starting count: 2,300 / 2,000 = 115 percent, which tells you about growth, not retention.

Converting between periods

Retention compounds. A monthly retention rate of 97 percent does not mean an annual rate of 97 percent, nor of 64 percent (which would be 100 percent minus twelve months of 3 percent churn). The annual rate is 0.97 raised to the twelfth power, which is about 0.694, or 69.4%. A cohort loses about 31 percent of its members in a year at that monthly rate.

Going the other way, an annual retention of 80 percent corresponds to a monthly rate of 0.80^(1/12) = 0.9816, or 98.2% (monthly churn of 1.8 percent). Always convert using powers, not multiplication, and always label whether a figure is monthly, quarterly or annual.

What "retained" means

For a subscription, a customer is retained if the subscription is active at the end of the period. For other businesses the definition needs a decision.

Contractual relationships (subscriptions, memberships, B2B contracts): active at period end.

Non-contractual relationships (retail, ecommerce, restaurants): customers do not announce that they have left. Use one of:

• Repeat purchase rate: the share of customers with at least two orders. If 1,000 customers made a first purchase in a year and 280 made at least one more, the repeat purchase rate is 28 percent.
• Returning customer rate over a window: the share of last year's customers who bought again in the following twelve months.
• Time-based definition: a customer is retained if they purchased within a window equal to about twice their usual repurchase interval. A replenishment product with a 60-day cycle might use a 120-day window.

The right window depends on the product. A company selling mattresses should not measure 90-day retention. A company selling coffee pods should not wait a year.

Retention and lifetime value

Retention rate drives expected lifetime. With annual retention of r, a simple estimate of expected lifetime is 1 / (1 - r) years.

Moving retention from 80 to 85 percent raises expected lifetime by a third, from 5.0 to 6.7 years. The same improvement from 70 to 75 percent raises it from 3.3 to 4.0 years, a 20 percent increase. Gains are larger when retention is already high, which is why mature businesses often find retention the cheapest source of growth. In practice, check any lifetime estimate against cohort evidence, because a projected ten-year lifetime may not hold. See customer lifetime value.

The economics of retention

Consider a business with 10,000 customers, annual retention of 80 percent and a goal of 5 percent customer growth. To grow from 10,000 to 10,500 while losing 20 percent (2,000 customers), it must acquire 2,500 new customers. At CAC of $80, that costs $200,000 a year. If retention improved to 85 percent, losses would fall to 1,500 and the company would need only 2,000 new customers to reach 10,500, costing $160,000. Five points of retention saved $40,000 of acquisition spend, or 20 percent of the budget, which could be reinvested in growth.

Ways to improve retention

Prioritize by finding where the cohort curve drops most. If most customers never place a second order, work on the second purchase. If customers drop after six months, look at the usage cycle and the product experience at that point.

Measuring retention programs honestly

When you launch a retention program, compare a treated group with a similar untreated group, using a holdout where possible. Otherwise, you may credit the program for repeat purchases that would have occurred anyway. Example: an email campaign sent to 10,000 customers produces 800 purchases, while a 1,000-customer holdout produces 60 purchases, a rate of 6 percent. The campaign group rate is 8 percent, so the incremental purchase rate is 2 points, or 200 incremental purchases from the 10,000 emailed, not 800. At $22 of contribution per order and campaign cost of $2,000, incremental contribution is 200 × 22 = $4,400, a net gain of $2,400. Crediting all 800 purchases would have shown $17,600 of contribution and overstated the benefit more than sevenfold.

Link to the supply chain

Many retention failures start upstream. A stockout on a replenishment item sends the customer to a competitor, late deliveries reduce satisfaction, and inconsistent quality drives returns. Reliable availability is a retention strategy. For consumer product companies, demand planning and inventory policy directly support customer retention. See demand planning and inventory optimization.

Common mistakes

• Including new customers in the numerator.
• Using the wrong time window for the product's purchase cycle.
• Mixing monthly and annual rates.
• Averaging across segments with different behavior.
• Crediting retention programs for purchases that would have happened anyway.
• Ignoring revenue retention when customers differ in size. See revenue churn vs. logo churn.

Retention by segment: a worked example

Company-wide retention hides the groups that matter. A hypothetical business has two customer groups over a year. Group P starts the year with 4,000 customers and keeps 3,400, so annual retention is 3,400 / 4,000 = 85%. Group Q starts with 6,000 customers and keeps 3,900, so annual retention is 3,900 / 6,000 = 65%. The company-wide figure is (3,400 + 3,900) / 10,000 = 73%, which describes neither group. If group Q is acquired through a discount-driven channel and group P through referrals and organic search, the finding changes the acquisition plan: paying more for customers who look like group P is justified, while paying the same price for group Q is not.

What to put on a retention dashboard

Keep the dashboard short. Show retention rate by month and by cohort, repeat purchase rate, time to second order, retention by channel and by first product, revenue retention for the top customers, and the number of customers at risk by your own definition. Each chart should have one owner who is expected to explain movements at the monthly review. A dashboard that nobody owns is simply decoration.

FAQs

What is customer retention rate?

The percentage of customers at the start of a period who are still customers at the end, excluding new customers acquired during the period.

How do you calculate customer retention rate?

Subtract new customers from the ending customer count, then divide by the starting customer count.

What is a good customer retention rate?

It depends on the business model and purchase cycle. Compare cohorts over time, and judge it by what it does to lifetime value and payback.

How do you measure retention for a non-subscription business?

Use repeat purchase rate or the share of customers who buy again within a window that fits the product's usual repurchase cycle.

How do you annualize a monthly retention rate?

Raise the monthly rate to the power of twelve. A monthly rate of 97 percent gives an annual rate of about 69 percent.