SaaS Benchmarks Are a Trap: The Median Belongs to a Company That Isn't Yours

You know the ritual... A number feels off, so you type "good SaaS NRR" into Google and wait. Up pops a figure, so you hold yours next to it and either relax or panic.

You didn't learn much - That benchmark is an average smeared across hundreds of companies at different stages, selling to different buyers, pricing in different ways. It describes a company that doesn't exist. SaaS benchmarks feel like answers but most of the time they're someone else's homework.

Where the number actually comes from

Pull back the curtain on any benchmark and you find the same thing. A survey went out. A few hundred companies replied. Someone took the middle value and published it as "the benchmark." Clean, quotable, and almost useless for you specifically.

Because the companies in that sample are nothing alike. One sells a $4 app to consumers. One sells a $400k platform to banks. One is six months old. One has been compounding for a decade. Mash their retention numbers together and the median you get belongs to none of them. It's the statistical equivalent of the average household having 1.8 children and a third of a dog.

You wouldn't run your pricing off that. So why set targets by it?

The numbers don't even agree with each other

SaaS benchmarks openly contradict each other - and nobody seems bothered.

One widely cited roundup puts median net revenue retention for private B2B SaaS at around 101%. A different dataset in the exact same write-up pegs the B2B SaaS median at 82%. Same metric. Same rough year. A twenty-point gap. Pick whichever makes you feel better, apparently.

So which is "good"? Neither. Both. The question is broken.

Founders come to me terrified because their NRR is below some number they read on a blog. I ask where the number came from and they have no idea. It was just sitting at the top of a search result looking official. Half the time the benchmark was measured on enterprise companies and they're selling to tiny startups, so of course the maths looks different. The number spooked them into a decision that had nothing to do with their actual business. Ian Naylor, Founder at SaaSToolkit.ai

The median is a company that doesn't exist

The deeper you dig, the worse the median looks. Break that same retention data down by who the company sells to and it splinters completely. Enterprise-focused SaaS clusters up near a 118% median. SMB-focused companies sit closer to 97%. Twenty-one points apart, and that's before you touch pricing model or company age.

Free-to-paid conversion tells the same story. ChartMogul found the median across products is about 8%, but that tidy little number hides companies converting at 2% and others clearing 25%. If yours is 6%, are you failing? No clue. Not from that number. It depends entirely on who you sell to and how you sell it, which the benchmark quietly flattened into mush.

The median is a company that doesn't exist, and we keep treating it like a rival. I had a client beating themselves up over a conversion rate that was perfectly healthy for their market, because some report said the average was higher. It turned out that report was mostly self-serve products with no sales team, and my client sold a complex tool that needed demos. Totally different! Their real competition was their own last quarter, and against that they were up nicely. The SaaS benchmark just made a good quarter feel like a failure. Becky Halls, Strategist at SaaSToolkit

What SaaS benchmarks are actually for

None of this means bin them. A benchmark is a compass, not a destination. It tells you roughly which way the ground slopes. It does not tell you where to stand.

Used well, a benchmark answers one honest question. Is this metric wildly out of line, like dramatically off, enough that I should go investigate? If your churn is triple every number you can find, fine, that's a signal worth chasing. But sitting a few points under some blended median is noise. That's not a problem. You can burn a whole quarter sweating a gap that only exists because you're in a different business than the people surveyed.

Treat a SaaS benchmark like a weather forecast for a city you don't live in," says Priya Menon, a SaaS metrics consultant who helps early-stage teams set targets. "It's vaguely useful for knowing the season. It's useless for deciding whether to bring a coat today. The teams that grow fastest stop asking 'what's good' and start asking 'are we better than we were.' That second question you can actually answer with your own data, and it's the only one that moves the business. Everything else is just anxiety with a percentage sign on it.

Measure against your own past self

So here's the swap worth making. Stop grading yourself against a stranger's median. Start grading yourself against last month's you.

That shift changes everything about how the numbers feel. Instead of chasing an external target that was never built for your business, you watch whether this month's cohort behaves better than last month's. Is activation climbing? Is week-four retention firmer than it was in spring? Those are real, answerable, and yours. It's also why pinning everything to a single number you've picked to steer by can backfire. A benchmark and a lone north-star metric fail the same way, if you forget what they're actually measuring.

Your own trend line is honest. It knows your buyers, your pricing, your product. It can't be gamed by a survey sample from companies you'll never meet. And it tells you the one thing a benchmark never can: whether the work you did this quarter actually worked.

That's the whole game. Not catching up to a median. Beating your own baseline, month after month, until the compounding does the heavy lifting for you.

This is where having your numbers in one place earns its keep. SaaSToolkit pulls behaviour, referral and revenue data from a single JavaScript snippet, so your baseline lives in one spot instead of scattered across four tools that each tell a slightly different story. You can see your own cohorts move over time, in context, without stitching exports together on a Sunday night. The median will still be out there. You just won't need it anymore.

Know your own numbers well enough and the benchmark stops being a judge. It becomes background noise. A rough bit of weather for a city you don't live in.

FAQ

What is a SaaS benchmark? It's a reference figure, usually a median or average for a metric like churn, net revenue retention or conversion, drawn from a survey of many SaaS companies. The idea is to show you what "typical" looks like. The catch is that it blends very different businesses together, so "typical" rarely matches your specific stage, pricing or market.

Are SaaS benchmarks useful at all? Yes, but only as a rough compass. They're good for spotting when a metric is dramatically out of line and worth investigating. They're bad as targets, because a few points above or below a blended median usually says more about who got surveyed than about how your business is actually doing.

Why do SaaS benchmarks disagree so much? Because each one surveys a different set of companies and measures things slightly differently. One report might lean enterprise, another self-serve, another a specific region or funding stage. That's how you get the same metric quoted as 82% in one place and 101% in another. Neither is wrong. They're just describing different crowds.

What's a good net revenue retention rate? It depends entirely on who you sell to. Enterprise-focused companies often sit well above 110%, while SMB-focused products can be healthy closer to 100% or even below. Rather than chasing a universal "good" number, compare your NRR to your own trend and to companies with a genuinely similar model.

What should I measure instead of benchmarks? Your own trajectory. Track whether each new cohort activates, retains and expands better than the last. That internal trend reflects your real buyers and pricing, can't be skewed by someone else's survey, and tells you directly whether your recent work is paying off.

How does SaaSToolkit help? SaaSToolkit captures behaviour, referral and revenue data from one JavaScript snippet, so your baseline sits in a single place instead of across multiple tools. That makes it easy to watch your own cohorts and metrics move over time, in context, so you can measure progress against your past self rather than a benchmark built from companies nothing like yours.