Benchmarking & External Comparisons
How to compare your numbers to other teams, rivals, and the wider industry without fooling yourself.
What you'll learn
- Tell internal, competitive, and industry benchmarks apart
- Read percentiles and quartiles with confidence
- Spot a benchmark that quietly compares the wrong things
A number on its own rarely means much. A 4% conversion rate could be a triumph or a disaster — you simply can’t tell until you have something to hold it against. Benchmarking is the practice of putting your numbers next to a reference point so you know whether you’re winning, losing, or sitting comfortably in the middle. The trick is choosing the right reference point, because a comparison against the wrong thing is worse than no comparison at all.
The three places a benchmark can come from
The first kind is the internal benchmark. You compare a number to your own history or to a sibling team. Last quarter’s churn was 4%; this quarter it’s 3%, so you improved. The east region closes deals in 30 days while the west takes 45, so the east has something worth copying. Internal benchmarks are the cleanest because you control how the data is collected — same definitions, same systems, same people doing the counting.
The second kind is the competitive benchmark. Here you compare yourself to a specific rival or a named handful of them. “Their app loads in 1.2 seconds; ours takes 3.” This is the most motivating comparison and the hardest to get honest data for, because competitors don’t hand you their internal numbers. You end up leaning on public filings, app-store stats, third-party trackers, and educated guesses.
The third kind is the industry benchmark. You compare yourself to a broad average across many companies in your sector — “the median SaaS company churns 5% a year.” These usually come from analyst reports and surveys. They’re useful for sanity checks but blunt, because “the industry” lumps together businesses that may look nothing like yours.
Three reference points, from the cleanest data on the left to the broadest on the right.
Percentiles, quartiles, and “best in class”
Industry benchmarks rarely come as a single number, because companies are spread out. Instead they come as percentiles. If you’re at the 75th percentile on customer satisfaction, you score higher than 75% of the companies measured. The median is the 50th percentile — the exact middle, with half above and half below. The median is often more honest than a plain average, because one giant outlier can drag an average up while leaving the median untouched.
You’ll also hear data split into quartiles — four equal buckets. The top quartile is the best-performing 25%, and that’s usually what people mean when they say a metric is “best in class.” It’s a far more useful target than “above average,” because being slightly above an average that includes a lot of struggling companies is a low bar. Aiming for the top quartile means aiming to be genuinely good, not merely typical.
Rule of thumb: “above average” can mean barely better than the worst; “top quartile” means genuinely good. Always ask which bar a benchmark is setting.
Apples to apples: normalization
The deepest benchmarking mistake is comparing things that aren’t comparable. Your support team handles 500 tickets a month and a rival handles 5,000 — does that make them better? Not if they have ten times your customers. Raw totals mislead whenever the things you’re comparing are different sizes.
The fix is normalization: divide by something that makes the comparison fair. Tickets per customer, revenue per employee, cost per unit, churn as a percentage rather than a count. Once both numbers are expressed on the same per-something basis, you’re finally comparing apples to apples. A small team and a huge one can sit side by side honestly only after you’ve normalized.
When a benchmark quietly lies
Even a well-normalized benchmark can mislead if the definitions don’t match. Maybe their “active user” means anyone who logged in this year, while yours means someone who logged in this week. Maybe their “revenue” includes refunds and yours doesn’t. The numbers look comparable and aren’t. Other classic traps: a sample that’s too small to trust, data that’s two years stale, or an “industry average” that secretly blends your sector with three unrelated ones. Before you trust any external number, ask how it was defined, when it was collected, and who exactly is in the sample.
Spot it: which benchmark?
Read each comparison and decide which type of benchmark it is, then tap a card to flip it.
Sort the benchmarks
Drag each item into the bucket it belongs to — or tap an item, then tap a bucket. Hit Check placement when you’re done.
Here's where each one goes:
- Our past performance → Internal — your own history, most reliable.
- A specific rival's public stats → Competitive — named competitor, hard to source but clear.
- Analyst report of sector average → Industry — broad comparison, useful but blunt.
- Sibling teams in same system → Internal — same definitions, same company, cleanest.
- App-store rankings against named competitors → Competitive — head-to-head with specific rivals.
- Survey of 500 companies in our sector → Industry — wide sample, industry-level insight.
Tip: drag with a mouse, or tap an item then tap a bucket on touch screens. Get one wrong and the answer key appears.
How to use it
When someone drops a benchmark into a meeting, slow it down with three questions: “Where did this come from — us, a competitor, or the whole industry?”, “Is it normalized, or are we comparing different-sized things?”, and “Does their definition match ours?” When you set a target, prefer “reach the top quartile” over “beat the average,” because it aims at real quality. And when you present your own numbers, bring the reference point with you: don’t just say “we’re at 4%,” say “we’re at 4%, which is the industry median, and top quartile is 6%.” That single sentence turns a bare number into a story your audience can actually judge.
Quick check
1. Comparing your team's results to last quarter's is which kind of benchmark?
2. Being at the 75th percentile means you score higher than…
3. "Normalization" mainly fixes the problem of…