Churn rate is the percentage of customers (or revenue) you lose over a given period. The basic formula is (Lost Customers ÷ Starting Customers) × 100, so losing 50 of 1,000 starting customers is a 5% churn rate. The details that trip people up are which denominator to use, whether you are measuring customers or revenue, and the fact that monthly churn compounds rather than multiplying by 12.
This page is the definition and the calculation mechanics. For typical rates by sector, what counts as good, and how to bring churn down, see the pillar guide on SaaS churn and retention benchmarks.
What is Churn Rate?
Churn rate is the percentage of customers who stop using your product or service during a specific time period. The standard mental model is a bucket with a hole in it: new customers flow in at the top, churned customers leak out of the bottom, and churn rate measures the size of the hole. It matters because it compounds. A business can add customers every single month and still shrink if the leak is bigger than the inflow.
The Churn Rate Formula
Churn Rate = (Lost Customers ÷ Starting Customers) × 100Customer churn rate
Worked example
You start January with 1,000 customers and 50 have cancelled by 31 January:
(50 ÷ 1,000) × 100 = 5%Monthly customer churn
Customer churn vs revenue churn
These are two different formulas answering two different questions, and mixing them up is the most common reporting error in early-stage reporting.
| Customer (logo) churn | Revenue churn | |
|---|---|---|
| Counts | Accounts lost | Recurring revenue lost |
| Formula | (Customers lost ÷ Starting customers) × 100 | (MRR lost ÷ Starting MRR) × 100 |
| Includes downgrades? | No, a downgrade is not a lost logo | Yes, contraction counts |
| Answers | Does the product work for the typical customer? | What is happening to the money? |
They diverge whenever your customers are not the same size, which is always. Lose ten small accounts out of 1,000 and customer churn is 1% while revenue churn might be 0.2%. Lose one enterprise account and the reverse happens. Report both, because either one alone can hide the problem.
Revenue Churn = (MRR Lost ÷ Starting MRR) × 100$4,000 lost from $100,000 starting MRR = 4%
Which denominator should you use?
This is the question that actually decides your number, and it is where most churn figures become uncomparable.
Take a month where you start with 1,000 customers, add 200 new ones, and lose 60. Three defensible-looking denominators give three different answers:
| Denominator | Calculation | Churn rate |
|---|---|---|
| Start-of-period customers (standard) | 60 ÷ 1,000 | 6.0% |
| Start plus new customers | 60 ÷ 1,200 | 5.0% |
| Average of start and end | 60 ÷ 1,070 | 5.6% |
Use start-of-period as your default. It is the common SaaS reporting convention, which means it is the version investors will benchmark you against, and it is the most conservative of the three. The other two flatter you by enlarging the denominator with customers who had almost no opportunity to churn yet.
The underlying rule, put well by Dave Kellogg, is consistency: whatever you include in the numerator must also appear in the denominator. If you count a mid-period signup who cancelled in the same month as churn, that customer has to be in the denominator too. If you exclude them from the denominator, exclude them from the numerator as well and let them show up in next period's figure.
Simple method vs cohort method
Both are useful and they answer different questions.
The simple method is the formula above, applied to your whole base each period. It is what you report monthly, it is fast, and it is what most benchmarks use.
The cohort method freezes a group of customers who started together and tracks what fraction remains over time. This is the honest way to see whether retention is improving, because it compares like with like. A blended monthly figure can look stable while each new cohort quietly retains worse than the one before.
Two rules for cohorts, both from Kellogg's work on churn measurement:
- Always run cohorts forwards. Start with the group as it existed on day one and follow it.
- Never calculate a retention rate in reverse. Working backwards from today's surviving customers produces survivor bias, because the accounts that left are missing from the starting group. The number comes out flattering and wrong.
Converting monthly churn to annual churn
Multiplying monthly churn by 12 is the single most common mistake in SaaS metrics. Churn compounds against a shrinking base: if you lose 3% in January, February's 3% applies to the 97% still there, not to the original 100%.
Annual Churn = 1 − (1 − Monthly Churn)12Compound the retention rate, then subtract from 1
| Monthly churn | Actual annual churn | Naive × 12 | Overstated by |
|---|---|---|---|
| 1% | 11.4% | 12% | 0.6 pts |
| 2% | 21.5% | 24% | 2.5 pts |
| 3% | 30.6% | 36% | 5.4 pts |
| 5% | 46.0% | 60% | 14 pts |
| 7% | 58.1% | 84% | 26 pts |
| 10% | 71.8% | 120% | impossible |
The error grows with the rate, and at 10% monthly the naive version returns 120%, which should be the giveaway: you cannot lose more than all of your customers. Note also that the comparison runs the other way when someone quotes you an annual figure. To go from annual back to monthly, take the twelfth root: monthly churn = 1 − (1 − annual churn)1/12.
Churn rate and average customer lifetime
One useful thing falls straight out of the churn number. The reciprocal of your churn rate estimates how long an average customer stays:
Average Customer Lifetime = 1 ÷ Monthly Churn Rate5% monthly churn implies a 20-month average lifetime
At 5% monthly churn the average customer lasts about 20 months; at 2% about 50 months. This is the figure that feeds lifetime value, which is why a small change in churn moves your unit economics so much. Treat it as an estimate rather than a fact, since it assumes a constant churn rate and real churn is front-loaded in the first few months.
Types of Churn
| Type | What it means | Example |
|---|---|---|
| Customer churn | The account stops paying entirely | Cancelling a subscription |
| Revenue churn | Revenue falls without the account leaving | Downgrading from premium to basic |
| Voluntary churn | The customer chooses to leave | Switching to a competitor |
| Involuntary churn | The customer leaves without deciding to | Failed payment, expired card, business closure |
| Gross churn | Revenue lost, ignoring any expansion | Used to size the leak itself |
| Net (or negative) churn | Churn after offsetting expansion revenue | Can go below zero if expansion exceeds losses |
Involuntary churn is worth separating out because it is the cheapest to fix. Those customers did not want to leave, so dunning, card-updater services and retry logic recover a meaningful share of them without changing anything about the product.
Common calculation mistakes
- Multiplying monthly churn by 12. Compound it instead, per the table above.
- Switching denominators between periods. Any trend you see is then partly an artefact of the method. Pick start-of-period and keep it.
- Mixing customer and revenue churn in the same conversation, or quoting whichever looks better.
- Calculating retention backwards from surviving customers, which builds in survivor bias.
- Blending segments. Self-serve and enterprise churn at completely different rates; a blended number tells you nothing actionable.
- Ignoring involuntary churn. If failed payments sit inside your headline number, you may be trying to fix a product problem that is really a billing problem.
- Counting a downgrade as a churned customer. That is contraction, and it belongs in revenue churn.
Churn Rate FAQ
How do you calculate churn rate?
Divide customers lost in a period by customers at the start, then multiply by 100: (Lost Customers ÷ Starting Customers) × 100. Losing 50 of 1,000 starting customers is a 5% churn rate.
What is the churn rate formula for revenue?
(MRR Lost ÷ Starting MRR) × 100, and unlike customer churn it includes downgrades as well as cancellations. Losing $4,000 of recurring revenue from a $100,000 starting base is 4% revenue churn.
Which denominator should I use for churn rate?
Start-of-period customers or MRR. It is the standard SaaS convention, so it keeps you comparable to published benchmarks, and it is the most conservative option. Whatever you choose, anything counted in the numerator must also be in the denominator.
How do you convert monthly churn to annual churn?
Compound it: annual churn = 1 − (1 − monthly churn)12. Do not multiply by 12. At 5% monthly the real annual figure is 46%, not 60%.
What is the difference between customer churn and revenue churn?
Customer churn counts accounts lost; revenue churn measures recurring revenue lost, including downgrades. A business can lose many small accounts while revenue barely moves, or keep nearly every logo and still lose serious revenue to one downgrade.
How do you calculate average customer lifetime from churn?
1 ÷ monthly churn rate, expressed in months. A 5% monthly churn rate implies roughly a 20-month average lifetime. It assumes churn stays constant, so treat it as an estimate.
What is a good churn rate?
It depends entirely on your model and segment, and the honest answer needs benchmarks rather than a single number. See SaaS churn and retention benchmarks for typical rates by industry and segment, plus what to do about a high one.
Related churn and retention metrics
- Churn rate, the definition and the calculation. (you are here)
- SaaS churn & retention, the pillar guide: benchmarks by industry, what good looks like, and how to reduce churn.
- Revenue churn rate, the money-weighted version in depth.
- Gross churn vs negative churn, before and after expansion revenue.
- Churn reduction techniques, the tactics for bringing it down.
- Net revenue retention, the same picture viewed from the retention side.
