What good retention actually looks like for B2B products

Every founder I work with eventually asks the same question. "Is our retention good?"

It is not answerable as asked. Retention only means something against a segment and a definition, and most teams are fuzzy on both. So here are the numbers, and then the part that matters more than the numbers.

First, which retention

Two different measures get called retention, and they answer different questions.

User retention is the percentage of users who signed up and are still active some period later. For B2B the useful window is six months. It answers: does the product keep being worth opening?

Net revenue retention compares the revenue you get today from a cohort of customers against the revenue you got from that same cohort a year ago. Expansion, contraction, and churn all land in one number. It answers: does an account grow or shrink once it is yours?

They can point in opposite directions, which is exactly why you need both. A product can lose half its individual users while revenue from those accounts climbs, because the buyer renewed and added seats that nobody uses. That is a real pattern in B2B, and it is a slow-motion churn risk dressed up as growth.

The benchmarks

Lenny Rachitsky ran the most useful version of this exercise I have seen, pulling benchmarks from more than twenty growth practitioners and investors into one set of numbers. For B2B, these are the ones worth writing down.

User retention after six months:

• SMB and mid-market SaaS: around 60% is good, around 80% is great
• Enterprise SaaS: around 75% is good, around 90% is great

Net revenue retention after twelve months:

• Land and expand, SMB and mid-market: around 90% is good, around 110% is great
• Bottom-up SaaS: around 100% is good, around 120% is great
• Enterprise SaaS: around 110% is good, around 130% is great

Read the enterprise row next to the SMB row and you can see why comparing yourself to the wrong segment is worse than having no benchmark at all. A 70% six-month retention is a decent result selling to small businesses and a warning sign selling to enterprises, where switching costs alone should hold most accounts in place.

What the number does not tell you

Here is where teams lose a quarter. The number comes back below the line, someone declares a retention initiative, and the backlog fills with re-engagement emails and an onboarding checklist. Six months later the number has not moved, because nobody ever established which users were leaving or why.

Retention is an outcome. Like every outcome, it moves when you find the specific cause underneath it, and it does not move when you attack it in general. That is the whole argument for treating metrics as a signal rather than a scoreboard: the number tells you where to look, and then you go and talk to people to find out why.

Three cuts usually do the work.

Cut by segment. An average retention number is a blend of a segment that loves the product and one that never should have signed up. If one segment retains at 80% and another at 30%, you do not have a retention problem. You have a targeting problem, and it is often solved by changing who you serve rather than what you build.

Cut by the first two weeks. Most B2B churn is decided long before the renewal. The account that lapses at month five usually never reached the moment where the product proved itself. That makes it an activation problem wearing a retention costume.

Talk to the ones who left. Churned users are the most informative interviews you can run and the ones teams avoid hardest, because the conversation is uncomfortable. Ask what they were trying to do, what they used instead, and when they stopped. Not what would have made them stay.

The pre-PMF caveat

If you are still looking for product-market fit, these benchmarks are not your test. They describe products that already have a market and are asking how well they hold it.

Before that point, the shape of the curve says more than the level. A retention curve that drops and then flattens means you found a group of people for whom the product is genuinely useful, even if that group is small. A curve that decays toward zero means you have not, no matter how good the first-week number looks. Flattening comes first. Raising the flat part comes after.

That is one of the harder things to hold onto when a board is asking for a percentage. It is also why I keep arguing that product-market fit is a hypothesis you test, not a milestone you announce.

How to use a benchmark well

Write down your segment first, then the two numbers, then the date. Compare against the row that matches your business, not the row you wish you were in. If you are below, resist the initiative and go find the cause. If you are above, check the other measure before celebrating, because strong revenue retention can be hiding weak usage for a year or more.

And do not let the benchmark become the goal. The goal is a product people keep choosing. Retention is just the most honest way we have of noticing whether they do. When the number and the strategy start disagreeing, that argument is usually worth having properly, which is most of what product strategy work turns out to be in practice.