A healthy acquisition funnel can hide a retention problem for months. New customers keep arriving, revenue appears stable, and then growth stalls because too many customers leave before they become profitable. Asking which metrics measure customer retention is the right starting point, but the best answer is not one dashboard number. It is a set of metrics that explains who stays, who leaves, when behavior changes, and what that movement means for revenue.
For business owners and marketers, the goal is not to report every available KPI. It is to choose a small group of measures that matches the business model and gives teams a reason to act.
Which Metrics Measure Customer Retention?
Customer retention is the share of customers who continue doing business with you over a defined period. The definition sounds simple, but the right measurement window depends on purchase behavior. A daily-use software product may track retention weekly or monthly. A furniture retailer may need quarterly or annual views. Measuring a customer as “lost” after 30 days makes little sense when the usual repurchase cycle is six months.
Start with the following core metrics. Together, they provide a useful view of customer health without turning reporting into an exercise in collecting numbers.
Customer retention rate
Customer retention rate, or CRR, is the foundational metric. It shows the percentage of existing customers you kept during a period, excluding newly acquired customers.
The formula is:
Customer Retention Rate = ((Customers at end of period – New customers acquired) / Customers at start of period) x 100
Suppose you begin a quarter with 1,000 customers, acquire 200 new customers, and finish with 1,050 customers. Your retention rate is 85%: ((1,050 – 200) / 1,000) x 100.
CRR works well because it separates loyalty from acquisition. Still, it does not show whether your retained customers are buying more, less, or at the same level. A subscription company can retain an account that has downgraded sharply, while a retailer can retain a customer who makes only a small repeat purchase. That is why CRR needs companion metrics.
Customer churn rate
Churn is the inverse perspective: the percentage of customers who stopped purchasing, canceled, or became inactive during a period. The basic formula is:
Customer Churn Rate = Customers lost during period / Customers at start of period x 100
For subscription businesses, churn is often straightforward because cancellation is a clear event. For ecommerce, professional services, and marketplaces, you must define inactivity. A customer may be considered churned after 90, 180, or 365 days without a purchase, depending on the normal buying cycle.
Track churn by customer segment, acquisition channel, plan, location, and tenure. An overall churn rate can look acceptable while a high-value cohort is leaving at an alarming pace. Also distinguish voluntary churn, such as a customer canceling, from involuntary churn caused by failed payments or expired cards. The fixes are very different.
Repeat purchase rate
Repeat purchase rate measures the percentage of customers who place more than one order within a chosen period. It is particularly useful for ecommerce, consumer brands, and transactional B2B companies where there is no recurring contract.
Repeat Purchase Rate = Customers with two or more purchases / Total customers x 100
This metric tells you whether customers return at all, but it does not reveal how quickly. Pair it with time to second purchase. If customers who return within 30 days tend to become long-term buyers, that first-month experience deserves concentrated attention. Post-purchase education, replenishment reminders, onboarding, or a well-timed service check-in may have more impact than another discount campaign.
Purchase frequency and time between purchases
Purchase frequency shows how often active customers buy during a period. Time between purchases shows the average gap from one transaction to the next. Both metrics help identify early warning signs before a customer meets your formal churn definition.
If repeat buyers are still returning but taking longer to do so, retention is weakening even if the retention rate has not yet moved. That pattern can point to pricing pressure, reduced product relevance, inventory issues, or a competitor gaining share.
Use median time between purchases alongside the average. A few customers with very long gaps can distort the average, while the median better reflects the typical customer experience.
Revenue retention: gross and net
Customer counts matter, but revenue retention often matters more. Gross revenue retention, or GRR, measures how much recurring revenue remains from an existing customer base after churn and downgrades, without counting expansion revenue.
GRR = (Starting recurring revenue – Churned revenue – Downgrade revenue) / Starting recurring revenue x 100
Net revenue retention, or NRR, adds upgrades, cross-sells, and expansions from those same customers.
NRR = (Starting recurring revenue – Churned revenue – Downgrade revenue + Expansion revenue) / Starting recurring revenue x 100
These are especially valuable for SaaS, agencies with recurring retainers, and membership businesses. An NRR above 100% means retained customers are generating more revenue than the cohort did at the start, even after losses. That can support efficient growth, but it should not excuse high logo churn. Losing many smaller customers may weaken referrals, brand perception, and future expansion opportunities.
Customer lifetime value
Customer lifetime value, or CLV, estimates the total gross profit or revenue a customer is likely to generate over the relationship. The appropriate calculation depends on the quality of your data. A simple transactional version is average order value multiplied by purchase frequency multiplied by estimated customer lifespan.
CLV becomes more useful when segmented. Compare the lifetime value of customers acquired through paid search, referrals, partner channels, or organic content. Then compare the cost to acquire each group. A channel with a higher acquisition cost may be the better investment if its customers retain longer, buy more frequently, and require less support.
Avoid treating CLV as a precise forecast. It is an estimate built on assumptions about future behavior. Use it as a directional decision tool and update it as cohorts mature.
Cohort retention
Cohort analysis groups customers by a shared starting point, usually the month or quarter they first purchased or subscribed. You then track what percentage remains active in each following period.
This view answers questions that aggregate retention cannot. Did customers acquired after a new onboarding flow retain better? Did a pricing change damage loyalty among smaller accounts? Are customers from a recent campaign less likely to return than those acquired through referrals?
A cohort table may look less tidy than a single retention percentage, but it is often the fastest way to find whether performance is improving or merely being masked by newer acquisition. For growing companies, this is one of the most decision-ready retention reports available.
Build a retention scorecard that prompts action
A practical scorecard usually combines one customer-count metric, one behavior metric, and one revenue metric. For example, an online store might monitor quarterly retention rate, repeat purchase rate, median time between orders, and CLV by acquisition source. A SaaS company may prioritize logo churn, GRR, NRR, product adoption, and failed-payment recovery.
The operating rule is simple: every metric should have an owner, a reporting cadence, and a plausible response when it changes. If repeat purchase rate falls, the lifecycle marketing team may test replenishment timing or post-purchase messages. If churn rises among customers who never use a core feature, the product and customer success teams should examine onboarding rather than defaulting to discounts.
Do not rely solely on satisfaction scores such as Net Promoter Score or customer satisfaction surveys. They can add context, especially when paired with open-text feedback, but stated intent and actual behavior often diverge. A customer may give a positive survey response and still leave because a competing option is cheaper, easier, or better suited to a changing need.
Set benchmarks carefully
There is no universal “good” retention rate. Industries, contract lengths, price points, and purchase cycles vary too much. A 90% annual retention rate may be excellent for one business and concerning for another. Your strongest benchmark is usually your own historical cohort performance, followed by comparable companies with similar models.
Look for trends before reacting to a single period. A one-month decline may reflect seasonality, a billing-calendar quirk, or a temporary stockout. A decline across several cohorts, segments, or months is evidence that deserves investigation. Review both percentages and customer counts, since small samples can produce dramatic percentage swings.
Retention measurement is valuable when it changes what the business does next. Choose metrics that expose the customer moments where value is created or lost, then use them to make those moments better. That is how a dashboard becomes a growth system rather than a monthly report.