Why I even bother comparing myself to a median
I run small, unglamorous online businesses. None of them are venture-backed, none of them have a board asking me to hit a growth curve. So when I first started pulling SaaS benchmark reports, I wasn’t trying to prove I was building the next big thing. I was trying to answer a much smaller question: is the number I’m staring at actually a problem, or is it just how this kind of business behaves?
That distinction matters more than it sounds. A 6% monthly churn rate feels alarming until you find out that self-serve, low-price SaaS products commonly sit somewhere in that range, while enterprise contracts with annual terms often sit much lower. Without a reference point, every metric is just a number floating in space. A benchmark gives it a neighborhood to live in.
This article is about how to use published SaaS benchmarks well, and just as importantly, how to avoid using them badly. I’ll walk through where these numbers come from, what “median” actually protects you from, and the specific ways I’ve seen people (myself included) misapply them.
Where benchmark numbers actually come from
Most public SaaS benchmark reports are aggregated from a pool of companies that opted into sharing data, usually through a survey, a billing platform’s aggregate dataset, or a VC firm’s portfolio review. A few things follow from that:
- The sample is self-selected. Companies willing to share churn and revenue data with a research firm are not a random slice of every SaaS business on earth.
- The sample skews toward a certain size and stage. A lot of these reports lean on venture-funded, growth-stage companies, because that’s who fills out the surveys and who VC-affiliated benchmark reports are built to serve.
- “SaaS” is a huge category. A benchmark blending together $9/month tools and $90,000/year enterprise platforms will produce a median that describes neither one particularly well.
None of this makes benchmarks useless. It just means the number you’re reading is a description of a specific population, not a law of nature. Before I compare my numbers to any published figure, I try to find out who was actually in the sample: what pricing tier, what customer segment, what stage of company. If a report doesn’t say, I treat the number as a rough compass heading, not a target.
Why median, and not average
If you’ve looked at more than one benchmark report, you’ve probably noticed they report the median instead of the mean. That’s a deliberate choice, and it’s worth understanding why.
Business metrics like churn, MRR growth, and customer acquisition cost tend to have long tails. Most companies cluster in a normal-ish range, but a handful of outliers, either doing exceptionally well or badly, can pull an average way off center. If ninety-eight companies in a sample have 40% annual churn and two have 400% annual churn because they’re liquidating, the average churn rate gets dragged upward in a way that doesn’t represent anybody in the group.
The median sidesteps that. It’s the middle value when you line every company up in order, so a handful of extreme outliers can’t distort it. That’s why median is the more honest number for “what does a typical company in this dataset look like.” It’s also why, when I see a report leading with an average instead of a median, I get a little more skeptical of the numbers.
What a benchmark can tell you
Used carefully, a benchmark is good for exactly one thing: telling you whether a number is unusual for your category. That’s genuinely useful. If your monthly churn is sitting at 15% and the reported median for your kind of product is closer to 3-5%, that’s a real signal worth investigating, even if the benchmark’s sample isn’t a perfect match for your business. Directionally, something looks off.
Benchmarks are also useful for catching your own blind spots. When you only look at your own numbers over time, you lose the sense of what’s normal versus what’s alarming, because you get used to whatever you’ve been living with. A median from outside your own bubble breaks that tunnel vision. It’s the same reason it helps to occasionally talk to another operator in your space, not to copy what they do, but to recalibrate your sense of “normal.”
What a benchmark can’t tell you
This is the part that gets skipped, and it’s the part I care about most.
A benchmark can’t tell you why your number is what it is. If your churn is above median, the report won’t tell you whether that’s because your onboarding is confusing, your pricing attracts the wrong customers, or you’re in a genuinely higher-churn category like a seasonal or trial-heavy product. You still have to go dig through cohort data, support tickets, and cancellation reasons to find the actual cause. The benchmark just tells you where to start looking.
It also can’t account for your specific business model. A one-person SaaS tool selling to solo freelancers is going to have a different churn profile than one selling multi-year contracts to mid-size companies, even if both get labeled “SaaS” in the same report. Comparing your solo-operator, self-serve product against a benchmark built mostly from funded, sales-led companies can make you chase a number that was never realistic for your situation to begin with.
And a benchmark definitely can’t tell you what’s “good enough” for your goals. Median just means half the sample is above it and half is below it. It says nothing about whether the median company is actually healthy, profitable, or sustainable. I’ve seen founders treat “beating the median” as the finish line, when the real question is always whether the business supports what you’re trying to build, at whatever pace that takes.
How I actually use benchmarks in practice
When I check a metric against a benchmark, I go through roughly the same steps every time:
- Find the closest matching segment. I look for a benchmark cut by pricing tier or business size that’s closest to mine, rather than grabbing the top-line number.
- Treat it as a range, not a line. Instead of asking “am I above or below the median,” I ask whether I’m in the same rough neighborhood, or clearly off in one direction.
- Use it to prioritize, not diagnose. If a number is well outside the typical range, that tells me where to spend my next hour of investigation. It doesn’t tell me what I’ll find.
- Recheck my own trend before comparing outward. A number moving in the wrong direction over the last three months matters more to me than where it sits relative to some external median, because my own trend has context a benchmark never will.
That last point is really the whole lesson. Benchmarks are a sanity check on your own data, not a replacement for looking at your own data closely. I still track my own churn, conversion, and retention cohort by cohort, month over month, because that’s the record that actually reflects my business. The median just helps me tell whether an unusual month is genuinely unusual, or just Tuesday for a company like mine.
If you’re building the habit of watching your own metrics closely, that’s really the foundation everything else here sits on. Benchmarks are a supplement to that habit, not a substitute for it.
Want more breakdowns like this on the metrics and methods behind running a business on your own numbers? Head back to the homepage for more.
Get new guides and videos first — join the Telegram channel.