Churn and retention: why traffic alone does not equal revenue

Traffic measures how many people arrive. Revenue depends on how many of them pay and how long they stay. A product with huge visits and high churn can earn less than a quiet product with loyal customers, which is why retention matters as much as reach.

This guide explains the key terms, summarizes recent benchmarks, and shows how to read traffic and revenue estimates with churn in mind.

Churn in plain words

Churn is the share of customers or revenue you lose over a period. If you start a month with 100 customers and 5 cancel, customer churn is 5 percent for that month.

There are several versions, and mixing them causes confusion.

• Customer churn counts the number of accounts lost.
• Revenue churn counts the money lost, which can differ if big customers leave.
• Voluntary churn happens when customers choose to cancel.
• Involuntary churn happens when a payment fails and the subscription lapses.

Retention is the opposite side. If churn is 5 percent a month, retention is 95 percent a month, but that compounds badly. Keeping 95 percent each month leaves about 54 percent of the original group after a year.

Why traffic misleads

A visit is not a customer. Between a visit and recurring revenue are at least four steps: signup, activation, payment and staying. Traffic only measures the first.

Consider two hypothetical products:

Product	Monthly visits	Customers who stay a year	Outcome
Tool A	500,000	Very few, mostly one-time users	Large audience, small recurring revenue
Tool B	20,000	Most, because it fits a daily workflow	Small audience, stable revenue

Tool A may look impressive on a traffic ranking, but Tool B may be the better business. The same is true when reading modelled revenue. The StartupsIntel estimate multiplies visits by a 3 percent paying share and the lowest paid plan, which assumes an average product, and it cannot see churn. Our methodology explains the limits.

What benchmarks say

Retention data is easier to find for B2B software than for AI consumer tools, so treat these as reference points.

SaaS Capital publishes retention benchmarks for private B2B companies. In the figures summarized there, companies with contract values between $25,000 and $50,000 show a median net revenue retention near 102 percent, with the top quartile around 111 percent and the lowest quartile around 97 percent.

Other 2025 summaries tell a similar story. One reports median net revenue retention near 102 percent and median gross revenue retention around 84 percent, down from the year before. Another puts median annual revenue churn for private B2B startups near 12.5 percent. Consumer and prosumer products fare worse, with some summaries citing gross churn of 30 to 45 percent a year.

Because these are secondary summaries and samples differ, use them as rough ranges.

One more claim you may see is that AI-native companies lose far more customers than traditional software, with a figure of 43 percent a year circulating. I could not trace it to a primary source, so I would not rely on it. Still, the direction is plausible for AI, where trying a tool is cheap and alternatives are plentiful.

Retention terms investors use

Three metrics show up in nearly every discussion.

• Gross revenue retention (GRR) is the share of starting revenue you keep after cancellations and downgrades, capped at 100 percent.
• Net revenue retention (NRR) adds upgrades and expansion. Above 100 percent means existing customers are paying more in total than a year ago.
• Cohort retention tracks a group of customers who joined in the same month and shows how many remain over time.

A business with NRR above 100 percent can grow even with no new customers. One below 100 percent has to keep refilling a leaking bucket.

For definitions of the revenue metrics themselves, see our MRR and ARR guide.

Why AI products have a retention challenge

AI tools face a few special pressures.

• Novelty effect. People try a new tool for a week and move on.
• Project-based use. A user may need one video or one logo, then stop.
• Low switching costs. Prompts and outputs can move to a competitor easily.
• Fast-moving models. A competitor with a better model can lure customers away.
• Free alternatives. Large labs keep adding capabilities to free tiers.

These forces suggest AI founders should watch churn closely and design for ongoing use, not a single burst.

What improves retention

You can improve staying power without heroic effort.

1. Get users to first value quickly. Retention is strongly tied to early success.
2. Build habits. Saved projects, templates, scheduled runs and integrations make products part of a routine.
3. Reduce involuntary churn. Retry failed payments, send card-expiry reminders and update billing details smoothly.
4. Offer annual plans. Fewer decision points mean fewer cancellations, as explained in monthly versus yearly plans.
5. Listen at cancellation. A one-question exit survey shows patterns.
6. Price for fit. Plans that match real usage reduce bill shock and regret.

How to judge a product from the outside

You cannot see a competitor's churn directly, but there are clues.

• Traffic trend. Steady or rising visits over many months suggest a lasting audience, while a spike and decline suggests novelty.
• Brand searches and direct visits. A healthy share of returning visitors points to loyalty.
• Review sites and communities. Complaints about cancellation or quality hint at churn.
• Pricing structure. Annual-heavy offers or usage credits can signal how the company manages retention.
• Age. Products that have survived several years have usually found some retention.

On StartupsIntel, the trending rankings show which products are climbing, and a site report shows tech stack, SEO health and domain age, which add context beyond visits. Treat all of it as signals, not proof.

When you are buying a business

If you are evaluating a startup to buy, churn is one of the first questions to ask. Request cohort data by month and revenue by customer. Check whether growth comes from new users replacing lost ones. Verified payment data matters here, which is why the Buy a startup marketplace emphasizes revenue verified from the payment provider.

A simple retention calculation you can do today

You need only two numbers per month: customers at the start and customers lost during the month.

If you began with 200 customers and lost 12, monthly customer churn is 6 percent and retention is 94 percent. Raise 0.94 to the twelfth power and you get about 0.48, which means roughly half of today's customers would remain after a year if nothing changed. That is a sobering way to see why a few points of monthly churn matter so much.

Now imagine you cut churn from 6 to 4 percent. Twelve-month survival rises from about 48 percent to about 61 percent. Small improvements in retention compound into much larger customer bases, which is why many founders find retention work pays back faster than buying more traffic.

FAQ

What is a good churn rate for a SaaS product?

It depends on the customer type. Enterprise software often sees low annual churn, while small-business and consumer products see much higher numbers. Compare yourself with products that sell to the same kind of buyer.

Is high traffic a sign of a successful startup?

Not by itself. Traffic shows interest, but success also requires conversion and retention. A smaller product with loyal paying customers may be healthier than a busy one with high churn.

What is net revenue retention?

Net revenue retention compares the revenue from a group of customers now with what they paid a year ago, including upgrades, downgrades and cancellations. A figure above 100 percent means the group pays more in total than before.

How can I estimate churn for a competitor?

You cannot measure it directly, but traffic trends, return visits, reviews and pricing structure give hints. For hard numbers you need access to their payment data.