Customer retention rate tells you how many existing customers you keep over a set period. The formula behind it takes three inputs and looks like this: ((E − N) ÷ S) × 100.
Most guides stop right there. But the formula is the easy part. What trips teams up is deciding what to count and over what window.
Say a company pays for 20 user seats. One team counts that as a single customer, while another counts all 20. Run the same formula on the same company, and those two teams land on different numbers.
That difference is why so many retention numbers can't be trusted.
This guide covers the factors that shape the result. You'll get the formula, a worked example, and current benchmarks with named sources. We'll also look at the mistakes that quietly push the number up and more.
Disclaimer: This content contains some affiliate links for which we will earn a commission (at no additional cost to you). This is to ensure that we can keep creating free content for you.
Customer retention rate is the share of existing customers still with you at the end of a period.
New customers don't count here. If you included them, you'd be measuring growth instead of retention. The whole point is to see how many of your current customers chose to stay.
It helps to be clear on what this number leaves out. It doesn't tell you how much those customers spent. That's a separate metric called revenue retention.
The same idea shows up under other names. Agencies often call it the “client retention rate.” SaaS teams talk about logo retention, which counts whole companies instead of seats. The math stays the same either way.
The reason it matters comes down to cost. Winning a customer takes real money, and you spend it again every time one leaves. Keeping the ones you have costs less, which is why customer retention strategies get lots of attention.
It's the share of customers you already had who stayed to the end of the period. New sign-ups are left out, and it says nothing about how much anyone spent.
Customer retention rate matters more than acquisition because the payoff keeps building over time.
Every new customer you chase is a fresh cost. But a customer you keep can pay you for years with no new cost to win them. That's why the best B2B marketing strategies optimize for retention.
The revenue data backs this up. SaaS Capital's 2026 survey looked at more than 1,000 private B2B SaaS companies. Their median net revenue retention was 103%:
Image via SaaS Capital
Anything above 100% means the existing base grows on its own before a single new sale.
Losing customers gets expensive in the other direction too. PwC's 2025 Customer Experience Survey asked 5,511 US shoppers about their habits. Roughly 52% stopped using or buying from a brand because they had a bad experience with its products or services, while nearly a third 29% stopped due to poor customer experience.
So next time you review your pipeline, give retention equal weight.
Retention keeps paying while acquisition starts over each month. A customer you hold onto earns more for less spend. PwC found 52% of shoppers drop a brand after one bad experience.
The customer retention rate formula is ((E − N) ÷ S) × 100.
Each letter stands for one count you pull from your own records. S is the number of customers you had at the start of the period. E is the number you had at the end. N is the new customers you gained in between.
Subtracting N is the whole reason the formula works. Leave those new customers in, and you're really measuring growth. That's a different thing from retention, even though people mix the two up often.
Multiply by 100 only at the very end. You'll track this number with your other sales funnel metrics, so it pays to get it right.
Watch out for shortcuts too. Some pages drop the N adjustment and just divide E by S. That version quietly flatters companies that are growing fast. Others swap in a different formula altogether, closer to a repeat purchase rate.
Each variable comes from a different part of your system. Each can also fail in its own way. S and E are snapshots taken on two fixed dates. N counts events that happen between those dates. That difference is why N causes the most trouble.
| Variable | What It Means | Where You Pull It | The Trap |
|---|---|---|---|
| S | Customers at the start of the period | A dated snapshot, not today's list | Rebuilding it from a current export quietly drops everyone who already churned |
| E | Customers at the end of the period | A dated snapshot at the close | Counting people still inside a cancellation notice as retained |
| N | New customers gained during the period | An event count between the two dates | Removing churned new customers from N, when the standard is gross |
Before you run the numbers, find out which systems can give you a dated snapshot. Some of them can't. Check whether your marketing reporting software can even export one cleanly.
S is your customer count at the start, E is the count at the end, and N is the new customers added in between. Only N is an event count across the whole period, which is why it causes the most reporting errors.
To calculate customer retention rate, apply the formula ((E − N) ÷ S) × 100.
In practice, that breaks into four steps. First, pick a period. Then pull three counts from one system. Subtract the new customers from your ending count. Divide by your starting count, and multiply by 100.
Finally, rebuild it in a spreadsheet to keep track of your work. Mistakes usually happen because of the setup and not the math.
Match your measurement period to your billing or renewal cycle. For example, if you measure annual contracts monthly, you’ll likely get a distorted reading that hides the actual performance.
Annual contracts usually run through subscription billing platforms — match that same cadence.
The period you pick shifts the result more than any other choice. For context, a 2% monthly churn doesn't add up to 24% over a year. It compounds to about 22% because each month works off a smaller base.
Pull all three counts from one source on the same dates. Say you take S from your CRM and N from billing — you'll end up with two different rates. Both will look correct, but they'll clash.
Your web analytics tools are rarely the right home for this. Pick one system of record, and write that choice down. For smaller teams, that record is usually in CRM tools for small businesses.
Let’s look at a customer retention rate example with real numbers for better perspective. A B2B software company starts the quarter with 500 accounts. It ends with 520. Along the way, it signed 60 new customers, so you adjust the ending count first.
Take E minus N, which gives you 520 − 60 = 460. Divide that by S, so 460 ÷ 500 = 0.92. Multiply by 100, and the rate comes to 92%.
Look closely at what happened. The account total rose by 20. Even so, 40 of the original 500 customers walked away.
If you want a customer retention rate calculator, four spreadsheet cells will do. Put S in cell A2, E in B2, and N in C2. Then type =((B2-C2)/A2)*100 into D2. Format the answer to one decimal place, and you're done.
That covers how to calculate customer retention rate in Excel or Google Sheets. Lock those input cells, and label the date range clearly.
Your customer retention rate can be wrong for three common reasons, including:
Each error is easy to miss. And each one leaves the number looking fine while it quietly misleads you.
Most reported figures are correct on paper but off in practice. The good news is that all three problems have clear fixes.
The standard rule counts N as all your new customers, not just the ones who stayed. For example, say you gain 30 customers and lose 12 in the same period. You still use 30 for N, not 18. So a customer who signs up in March and cancels in April stays in the count.
That same cancellation also lowers E — your ending count. So the math ends up punishing growth. Take a company adding lots of new customers, with some early churn. It can understate how well it keeps its core base.
You can fix this by weighting each customer by the time they were present. At the very least, run both the gross and net numbers, then report the gap.
A period that cuts across your renewal cycle optimizes for a question you never asked. Say most contracts renew in January. Calendar Q1 and fiscal Q1 will then report different rates from the same data. Neither one’s wrong, and that's what makes it tricky.
Small customer counts cause a related problem. With only 20 customers, a single cancellation moves the rate five points. Most marketing analytics tools show you the percentage, not the raw base.
Report the raw counts next to the percentage until your base grows large enough.
Definitions tend to drift, and your rate can climb while the business stays flat. A team might drop “non-target” churn as out of scope. Another might reset the base after a data migration.
None of this is dishonest. Each change is a fair call, made once and rarely looked at again. Still, the effect adds up and quietly lifts the number.
Three quiet errors usually throw it off. They involve how you treat N, which period you pick, and how the definition drifts over time.
A customer is whatever unit you decide to count, so you must define it before you calculate.
That single choice shifts your result more than any part of the formula. You might count accounts, seats, whole companies, or individual buyers.
Pick one of these units, then keep it steady for the whole year. An account is the contracting entity. A seat is one licensed user. A logo is a whole company, and a buyer is anyone who made a purchase.
The same business gives four different percentages, depending on which you pick.
Say a customer drops half its seats but keeps the contract. You keep that logo in full, but only half of those seats.
There are a few edge cases that can swing your figure more than the formula does. Leave out free trials unless they converted during the period, since counting them inflates S. Remove refunded customers from S, and from E too if they were still counted there.
Dormant accounts are trickier to handle. Someone inactive for eight months, but not canceled, still counts as a customer to billing. But your success team sees the same person as a loss.
Failed payments create involuntary churn, which your subscription management software records for you. Good recurring billing software usually flags these failed charges early too.
When you have no subscriber list, you build a denominator from purchase history instead.
A non-contractual business, like an ecommerce store, has buyers rather than subscribers. No one formally cancels, so people just stop coming back. That leaves the standard formula with no starting count.
The fix is a lookback window. Count everyone who bought in the past twelve months, then check how many returned through your ecommerce conversion funnel. Winning those buyers back is really how you increase ecommerce sales over the long run.
Clean that list for duplicates first. One buyer can show up as several as cookie tracking fades.
Rebuild your buyer list on a rolling twelve-month window each time you measure.
A customer is whatever unit you define first, such as an account, seat, logo, or buyer. Non-contractual businesses count buyers in a rolling twelve-month window instead.
A good customer retention rate meets or beats the median for your sector and contract size. Measure it over the same period as the benchmark you're comparing against.
No single number works for everyone. Across industries, a good customer retention rate can swing by more than forty points.
Before you trust any benchmark, find out where the number came from. Many “retention rate by industry” tables copy the same figures from each other. You'll see the same industry percentages repeated across guide after guide.
Take the widely quoted B2B retention rates, like “86% for software” or “88% for IT services.” Most of them trace back to CustomerGauge's B2B benchmark. That research is useful, but its data was collected between 2019 and 2021. Pages still republish it today with fresh-looking dates.
So an old survey isn't a current benchmark. Click through to the source before you rely on any figure. Then track the fresh numbers in your own marketing analytics tools.
Every figure below comes from a named source with a publication year. Each row also states its period so that you can compare fairly.
Remember that a churn figure and a retention figure aren't the same thing. And you can't read an annual rate against a monthly one.
| Sector | Figure | Period | What It Measures | Source (Year) |
|---|---|---|---|---|
| SaaS/software | 3.22% | Annual | Median churn, Recurly network | Recurly (2026) |
| Ecommerce | 4.25% | Annual | Median churn, Recurly network | Recurly (2026) |
| B2B SaaS | 82% | Annual | Median net revenue retention | ChartMogul (2025) |
| B2C SaaS | 49% | Annual | Median net revenue retention | ChartMogul (2025) |
| Bootstrapped SaaS, $3–20M ARR | 103% | Annual | Median net revenue retention | SaaS Capital (2026) |
| Premium SVOD, US | 4.6% | Monthly | Weighted-average churn, 2025 | Antenna (2026) |
| Telecom, postpaid phone | 0.89% | Monthly | AT&T postpaid phone churn, Q1 2026 | AT&T (2026) |
Two of these rows are monthly, not annual, so don't stack them against the yearly figures. And the two SaaS retention rows measure revenue, not customer headcount.
Recurly also gives you rough thresholds to judge annual churn. Below 2% is strong for almost any segment—most industry medians land between 3% and 5%. Once you climb above 5%, it's worth digging into why.
To benchmark like-for-like, match four things before you compare your rate to anyone else's. Line up the sector, the business model, the contract size, and the period.
Miss even one, and the comparison tells you nothing useful. The same care applies to every metric in your ecommerce KPIs.
SaaS Capital's 2026 report notes that comparing private companies to public SaaS has limited value. So find your own contract-value band first, then compare within it.
No universal figure exists, so a good customer retention rate beats your sector's median. Recurly's 2026 data treats annual churn below 2% as strong.
Customer retention rate and churn rate track the same base from opposite directions. Retention plus churn always adds up to 100% on the same base. Both also carry the gross-N effect we covered earlier. The two only split apart once you bring revenue into the picture.
What sets the other metrics apart is the unit each one counts.
Logo retention counts whole companies. Gross revenue retention and net revenue retention count money instead of customers. Repeat purchase rate counts transactions, and some tools blur it together with retention rate.
Customer lifetime value comes further down the chain. It tells you what a kept customer is worth, not whether you kept them. Cohort retention is the one extra metric worth tracking. It follows a single group of customers over time, rather than blending everyone.
A blended rate can look healthy at 85% while your newest customers quietly slip away. Your SaaS billing software can usually build that cohort view for you.
| Metric | What It Measures | Formula | When to Use It | What It Hides |
|---|---|---|---|---|
| Customer Retention Rate | Share of existing customers kept | ((E − N) ÷ S) × 100 | Logo-level health | Spend per customer |
| Customer Churn Rate | The share you lost | 100% − retention rate | Loss framing, same base | New-logo churn tucked inside it |
| Gross Revenue Retention | Revenue kept, no expansion | (Start MRR − churn − downgrades) ÷ start MRR | Worst-case revenue floor | Upsell performance |
| Net Revenue Retention | Revenue kept, with expansion | (Start MRR − churn − downgrades + expansion) ÷ start MRR | Whether the base self-funds growth | Logo losses masked by upsell |
| Repeat Purchase Rate | Share of buyers who bought again | Repeat buyers ÷ all buyers | Non-contractual ecommerce | Contract renewal behavior |
| Cohort Retention | Survival of one intake group | Cohort still active ÷ cohort size | Whether newer customers behave worse | Blended-average trends |
Real reports show why the labels matter. For instance, T-Mobile's Q1 2026 results list postpaid account churn of 1.04%:
Image via T-Mobile
A bigger shift is underway as well. As usage-based pricing spreads, logo retention and revenue retention drift apart. So add a revenue-based measure to your next report.
At the customer level, yes. Retention plus churn always equals 100% on the same base. At the revenue level, they split, since expansion can push net revenue retention above 100%.
Customer retention rate gets tracked almost nowhere by default. Most platforms store the raw inputs and leave the actual calculation to you. So measuring your customer retention rate is really a reporting job and not a tooling one.
Build the report inside whichever system already owns your customer record. For many teams, that's a CRM built for SaaS companies.
Moving data somewhere else just to compute one metric brings back the reconciliation problem. It also creates a second place for your definition to drift.
Others sync the fields across systems with CRM automation tools, while support-heavy teams feed them from customer portal software.
| System | Fields You Need | What It Gets Right | What It Can't Do |
|---|---|---|---|
| CRM (HubSpot CRM) | Lifecycle stage, close and churn dates, account hierarchy | Customer history in one record | Rarely computes retention on its own |
| Subscription billing | Subscription start, cancel, upgrade, downgrade | Accurate dates and revenue movement | Weak on non-paying relationships |
| Warehouse and BI | Dated snapshots of the customer table | Cohorts, segments, and full history | Needs modeling before it answers |
| Product analytics | Usage events per account | Early signals before a cancellation | Usage isn't the same as a contract |
Ecommerce teams often centralize this inside an ecommerce-focused CRM, while lean sales teams may prefer a CRM for lead management. If you can't recompute a rate later, it isn't auditable. The same goes for the data inside your customer feedback tools.
Calculate customer retention rate inside whichever system already houses your customer record. Use dated snapshots so past periods stay recomputable. Most CRM tools store the inputs and not the finished metric.
HubSpot Service Hub is the product in HubSpot's lineup built for retention work. Its Customer Success Workspace stores health scores and lifecycle data:
Image via HubSpot
It also runs net promoter score and customer satisfaction surveys. That gives you a strong layer of inputs to work from.
The catch is that Service Hub has no native retention report and no built-in cohort report. You build those yourself in its custom report builder, which starts on the Professional plan. The Free and Starter plans can't do it.
HubSpot Service Hub’s pricing is structured this way:
Image via HubSpot
None of this makes Service Hub a weak choice. Treat it as the store for your raw inputs, then build the customer retention rate report on top. Once you've selected your preferred system of record, start taking dated snapshots.
It stores your retention inputs, like health scores and NPS surveys, but computes no retention report itself. You build that in its custom report builder on Professional or higher.
To improve your customer retention rate, fix the point where customers drop off and set up an indicator that flags churn early. Good customer retention software helps with the second effort.
Customer churn usually gets decided long before anyone cancels. It happens early, in that window where a customer either reaches real value or not. Any win-back campaign after that point is merely fighting against an already made decision.
So onboarding is the cheapest place to lift your customer retention rate.
Map out where new customers stall, then fix the single worst step. Loyalty programs help, but they work on top of good onboarding, not instead of it. Layer proven customer retention strategies on top once that base is solid.
Deloitte's 2025 loyalty survey found that 72% of consumers spend more with a program. Those members were already onboarded first.
Your customer retention rate is a lagging metric, which limits what it can do. By the time the number moves, those customers have already left. You need a signal that shifts earlier. It should show you where to look while an account is still open.
Pick one leading indicator that tends to predict cancellation. That could be falling logins, a growing backlog in your IT ticketing systems, or a missed onboarding milestone.
Set a clear threshold, give one person ownership, and review it on a weekly cadence. An indicator that nobody owns won't get acted on. Some AI tools for customer success can surface these signals automatically.
Fix the earliest drop-off point first, which is usually onboarding. Then set up one leading indicator so at-risk accounts show up while you can still act.
Q1. Is a 90% Retention Rate Good?
A. In B2B software, 90% is usually a strong result. Still, “good” depends on your business model and contract size. A 90% customer retention rate can look great in one sector and weak in another. Compare it against your own sector's median, measured over the same period.
Q2. What Does an 80% Retention Rate Mean?
A. An 80% customer retention rate means 80% of the customers you started with stayed to the end. New customers are excluded from that count. The other 20% covers everyone who left, including new sign-ups who churned inside the period.
Q3. What’s a Good Customer Retention Rate?
A. No universal figure exists for a good customer retention rate. So judge yours against your own last four quarters first. Then compare with peers that share your contract size and renewal cycle.
Q4. How Do You Calculate Customer Retention Rate in Excel?
A. You calculate customer retention rate in Excel with three inputs and one formula. Put S, E, and N in cells A2, B2, and C2. Then enter *=((B2-C2)/A2)*100 in D2, and format it to one decimal place.
Q5. What’s the 80/20 Rule in Customer Retention?
A. The 80/20 rule suggests that a small share of customers are responsible for most of your revenue. As such, keeping those top customers matters more than the raw headcount suggests.
Q6. Can Customer Retention Rate Be Over 100%?
A. No. Customer retention rate can't go above 100%. You can't keep more existing customers than you started with, so it always caps at 100%. But net revenue retention can exceed 100%, since expansion revenue from staying customers offsets losses.
Q7. Does N Include New Customers Who Churned During the Period?
A. Yes — N counts all new customers, even the ones who canceled in the same period. So someone who signed and left still stays in N. Their exit also lowers E, your ending count. This drags down your customer retention rate for fast-growing companies. Time-weighting is the usual fix for this.
Q8. Why Do My CRM and Billing System Report Different Customer Retention Rates?
A. They differ because both systems define a customer and time events differently. For billing, subscriptions count from the payment date. For CRM, they count from the close date. Pick one system as your source of truth and reconcile the other against it.
You get your customer retention rate right by controlling the inputs, not just the formula. It's the share of existing customers who stayed, found with ((E − N) ÷ S) × 100.
The formula itself is a given. What decides your number is your inputs, your period, and how you define a customer.
So pick your unit, write that definition down, and measure it the same way every time.
Do that, and your customer retention rate becomes a number you can trust and act on. If you'd rather not build that health tracking yourself, the right tool keeps those signals in one place.
See how HubSpot Service Hub tracks customer health
Disclaimer: This content contains some affiliate links for which we will earn a commission (at no additional cost to you). This is to ensure that we can keep creating free content for you.
Customer onboarding software guides new clients through your product so they get value fast. It…
AI data analysis tools use AI and machine learning to examine data and identify patterns,…
Sales productivity tools are software applications that eliminate non-selling tasks such as CRM data entry,…
Generative AI for business creates content like text, code, images, and reports from instructions. It…
Marketing ROI is the profit generated by a marketing activity in relation to its cost.…
Email personalization uses customer data to create emails that feel relevant to each subscriber. It…