SaaS Retention Rate Calculation: 5 Formulas (2026)

Retention looks like one number until two teams calculate it and get different answers. Finance may measure retained recurring revenue, customer success may count renewed accounts, and product may track whether a signup cohort is still active.

This guide gives you the five formulas that matter, shows what each one answers, and uses the same example period throughout. You can copy the calculations into a spreadsheet or run them against your subscription database.

The quick answer

For the standard SaaS retention rate calculation, subtract new customers acquired during the period from ending customers, divide by starting customers, and multiply by 100.

Customer retention rate = ((customers at end - new customers) / customers at start) x 100

Example: you start the quarter with 400 customers, add 70, and finish with 430. Retained customers are 430 minus 70, or 360. Your customer retention rate is 360 divided by 400: 90%.

That is the right headline formula when you need to answer, ‘What share of existing customers stayed?’ It is not enough for every decision. A lost ₹1 lakh account and a lost ₹1,000 account both count as one customer, so revenue retention needs separate calculations.

Define the period and population first

Pick a fixed monthly, quarterly, or annual window before touching the numbers. Then decide which accounts qualify: paying customers, active subscriptions, eligible contracts, or activated product accounts.

Document these rules beside the metric. Exclude internal, test, refunded, and duplicate accounts consistently. Decide how you treat pauses, reactivations, upgrades, downgrades, and customers with multiple subscriptions.

The denominator must be the starting cohort. Adding new customers to the denominator makes acquisition look like retention and hides churn. This is the most common reason a retention dashboard appears healthier than the business actually is.

1. Customer retention rate

Customer retention rate, also called account retention, measures the share of customers present at the start who remain at the end.

Formula: ((ending customers - new customers) / starting customers) x 100

Using the example above, 400 starting customers and 360 retained customers produces 90% retention. The corresponding customer churn rate is 10%, because retention plus churn equals 100% when every starting customer is classified as retained or lost.

Use this metric for broad customer-base health and for comparing similar segments over time. Do not use it alone to forecast revenue. It weights every account equally and ignores upgrades and downgrades.

2. Logo renewal rate

Logo renewal rate answers a narrower question: of the contracts eligible to renew during the period, how many renewed? It is useful for annual contracts because many customers are not eligible to make a renewal decision in any given month.

Formula: renewed contracts / contracts eligible for renewal x 100

If 100 contracts come up for renewal and 84 renew, the logo renewal rate is 84%. Do not divide by every active customer. That would include contracts that had no chance to renew and would inflate the result.

Track the result by contract size, plan, tenure, industry, and customer success owner. A blended rate can look stable while a valuable segment deteriorates.

3. Gross revenue retention (GRR)

GRR measures how much recurring revenue from the starting customer base remains after churn and contraction. It ignores expansion, so it shows the durability of existing revenue before upsells soften the damage.

Formula: (starting MRR - churned MRR - contraction MRR) / starting MRR x 100

Suppose starting monthly recurring revenue is ₹10 lakh. You lose ₹60,000 from cancelled accounts and ₹40,000 from downgrades. GRR is (₹10 lakh - ₹1 lakh) divided by ₹10 lakh, or 90%.

GRR cannot exceed 100%. If your report shows 104%, expansion revenue has slipped into the calculation or the starting cohort changed during the period.

4. Net revenue retention (NRR)

NRR measures recurring revenue retained from the starting customer cohort after churn, contraction, and expansion. It tells you whether existing customers collectively shrink or grow.

Formula: (starting MRR - churned MRR - contraction MRR + expansion MRR) / starting MRR x 100

Continue the GRR example and add ₹1.5 lakh of expansion MRR from the same starting customers. NRR becomes (₹10 lakh - ₹60,000 - ₹40,000 + ₹1.5 lakh) divided by ₹10 lakh, or 105%.

An NRR above 100% means expansion from retained customers outweighed churn and contraction in that cohort. Still inspect GRR beside it. Strong upsells can mask a weak retention experience.

5. Cohort retention rate

Cohort retention tracks whether customers who started in the same period remain active after a defined interval. It is better than a blended average for diagnosing onboarding, activation, and product changes.

Formula: active accounts from the original cohort at interval N / accounts in the original cohort x 100

If 250 accounts activate in January and 175 are still active in month three, month-three cohort retention is 70%. Repeat the calculation for every signup month to create a retention table.

Define ‘active’ with a behavior tied to value, such as completing a core workflow, processing a transaction, or viewing a live dashboard. A login is easy to count but may not prove that a customer received value.

Which retention metric should you use?

Use customer retention when the question is how many customers stayed. Use logo renewal rate when you manage contracts with explicit renewal dates. Use GRR when you need downside visibility, NRR when you need the full revenue effect, and cohort retention when you need to diagnose product behavior.

A useful executive scorecard usually shows customer retention, GRR, and NRR together. Product and customer success teams should add cohort retention and renewal slices underneath. One universal retention number is tidy, but it is rarely honest.

A practical monthly calculation workflow

1. Snapshot all paying accounts and recurring revenue at the first second of the month. Store account ID, subscription ID, plan, segment, owner, and starting MRR.

2. At month end, classify each starting account as retained, churned, contracted, expanded, or unchanged. Keep newly acquired accounts separate.

3. Calculate customer retention, GRR, and NRR from the frozen starting cohort. Reconcile revenue movements to billing records before publishing the dashboard.

4. Segment the results. At minimum, compare plan, company size, customer age, and acquisition channel. A 92% blended rate can hide 70% retention in a new segment.

5. Attach an action to each threshold. Alert the customer success owner when a high-value account becomes inactive, notify finance when GRR drops, or send a webhook when a cohort falls below target.

Common SaaS retention calculation mistakes

Counting new customers as retained customers. New logos belong to acquisition. Remove them from the ending count before dividing by the starting population.

Mixing customers and subscriptions. One customer may hold several subscriptions. Choose an entity and use the same entity in the numerator and denominator.

Letting the cohort change. If you rebuild the starting population from today’s subscription table, deleted or merged records can rewrite history. Persist monthly snapshots or reconstruct them from an immutable event log.

Ignoring contraction. A customer who stays but cuts spend is retained for logo retention and lost revenue for GRR. Both facts matter.

Comparing unlike windows. Monthly retention cannot be compared directly with annual renewal rate. Label the interval in the metric name and dashboard.

Using averages without segments. Enterprise and self-serve customers have different buying cycles and retention patterns. Blend them only for a company-level view, then show the underlying segments.

Calculate retention from your database without manual exports

A spreadsheet works for an initial check. It becomes fragile when subscription events, invoices, product activity, and account ownership live in different tables. Manual exports also freeze the metric until someone repeats the work.

With AI for Database, you can connect PostgreSQL, MySQL, Supabase, BigQuery, MongoDB, and other supported databases with read-only access. Ask for customer retention, GRR, NRR, or a month-three cohort table in plain English, then inspect the generated result before saving it.

Turn the approved query into a self-refreshing dashboard so every team uses the same cohort definition. Add a workflow to send an email, Slack message, or webhook when retention crosses a threshold or an important account shows a risk signal.

Start with one metric and one segment. Validate it against a known billing period, document the rules, and only then automate it. That keeps speed from becoming confidently wrong arithmetic.

Questions SaaS teams ask

What is the simplest way to calculate SaaS retention? Use ((ending customers - new customers) / starting customers) x 100. Freeze the starting cohort and keep new acquisitions out of the numerator.

Should a SaaS team track GRR or NRR? Track both. GRR exposes churn and contraction without expansion; NRR shows whether expansion offsets those losses.

Can retention be above 100%? Customer retention and GRR cannot. NRR can exceed 100% when expansion revenue from the starting cohort is greater than churn and contraction.

Can I calculate SaaS retention without SQL? Yes. You can export billing data to a spreadsheet or use a natural-language database tool. Validate account definitions, time windows, and revenue classifications before relying on the result.

Final checklist

Before you share the number, confirm the time window, starting cohort, entity type, treatment of new customers, churn and downgrade rules, and segment filters. Show customer retention, GRR, and NRR side by side when revenue decisions depend on the result.

The formula is the easy part. Consistent cohort definitions and automated, reviewable data are what make retention useful enough to act on.