You ever try to average two percentages and realize the "obvious" way gives you a number that feels... off? Like, you got 80% on one test and 60% on another, so the average is 70%, right? Except when those tests weren't the same size, that 70% can quietly lie to you.
Here's the thing — averaging percentages trips up more people than you'd think. It shows up in grades, sales reports, survey results, ad click rates, and even cooking recipes if you're weird like me. And most spreadsheet tutorials treat it like a one-line formula, which is exactly why folks get burned.
What Is Averaging Two Percentages
So what are we actually doing when we say "average two percentages"? At face value, it sounds like adding them and dividing by two. And sometimes that's fine. But a percentage* is never just a number — it's a ratio. It's a part compared to a whole. 80% means 80 out of some 100, or 8 out of 10, or 800 out of 1,000. The size of that whole matters.
When you average two percentages, you're really asking: what's the combined rate across both situations? If they didn't, you need a weighted* average. If both percentages came from the same-sized base, a straight average works. That's the part most people miss.
Simple (Unweighted) Average Of Percentages
This is the 80% plus 60% divided by 2 version. On top of that, it's quick. You take the two numbers, add them, split the sum. It's clean. And it's only honest when the two percentages describe equal-sized groups or equal totals.
Say you run two polls with 100 people each. Which means one shows 70% support, the other 50%. But the unweighted average is 60%. That's real, because each poll carried equal weight.
Weighted Average Of Percentages
Now imagine poll A had 1,000 people and poll B had 100. A straight average still says 60%. But the bigger group should count more — it represents ten times the voices. On the flip side, the weighted version multiplies each percentage by its base size, adds those up, then divides by the total base. That's how you get the true combined rate.
Look, this isn't math snobbery. It's the difference between saying "about half the town agrees" and accidentally saying "most of the town agrees" when really 90% of your data came from one neighborhood.
Why It Matters
Why does this matter? Because most people skip the weighting step and report a number that looks authoritative but isn't.
In business, someone might average a 20% conversion rate from a tiny email test (50 visitors) with a 2% rate from a big campaign (5,000 visitors). Practically speaking, 1%. The naive average says 11%. Still, the weighted truth says about 2. Tell your boss the wrong one and you've got a strategy built on a rounding error.
In school, teachers who average exam scores without weighting by credit hours can bump a student's grade up or down by a full letter. In public health, averaging infection rates across counties of different population sizes makes small rural spikes look meaningless next to city data — or vice versa.
Turns out, the method you pick changes the story. And once a percentage gets into a slide deck, nobody questions how it was averaged.
How To Average Two Percentages
Alright, let's get into the actual doing. Also, if yes, simple average. Also, if no, weight them. The short version is: figure out if your percentages are from equal bases. Here's how each works, step by step.
Step 1: Identify The Base Behind Each Percentage
Before you touch a calculator, write down what each percentage came from. Percentage A = 75%, from 400 responses. Percentage B = 40%, from 100 responses. Done. Now you know they aren't equal.
If you don't know the base, you can't weight it. And real talk — if a report gives you only the percentages and says "average these," that report is incomplete. Ask for the raw counts.
Step 2: The Simple Average (Equal Bases)
If both bases are the same, or you're explicitly told to ignore size:
- Add the two percentages: A + B
- Divide by 2
Example: 90% and 70%. That's your average. Which means (90 + 70) / 2 = 80%. Easy.
Step 3: The Weighted Average (Unequal Bases)
This is the one you'll usually want. Formula in plain English:
- Convert each percentage to its count: A% × base_A, B% × base_B
- Add those counts together
- Divide by the total base (base_A + base_B)
- That result is your averaged percentage
Example: 75% of 400 = 300.And total count = 340. Total base = 500.40% of 100 = 40. That said, 5% (which is the simple average). Plus, not 57. 340 / 500 = 68%. Big difference.
Step 4: Watch Out For Percentage Points Vs Percentages
Here's a sneaky one. Think about it: know which one you're handling. And 5% only if bases match). That's different from averaging 10% and 15% as rates (which is 12.5 pts). If someone says "rate went from 10% to 15%, average the change," they might mean average the percentage points* (12.I know it sounds simple — but it's easy to miss in a hurry.
Continue exploring with our guides on what is a context clue definition and what is the difference between endocytosis and exocytosis.
Step 5: Sanity Check The Result
Does your averaged percentage sit between the two original numbers? It should. Also check: is the result closer to the bigger-base percentage? In weighted averaging, it must be. Even so, if you get 95% from averaging 80% and 60%, you typed something wrong. If 75% from 400 and 40% from 100 averaged to 45%, something's backwards.
Common Mistakes
Honestly, this is the part most guides get wrong — they list the formula and bail. The mistakes are where the real learning is.
Mistake 1: Averaging percentages from different base sizes without weighting. The classic. Looks fine, lies quietly.
Mistake 2: Averaging percentage changes instead of rates. If campaign A grew 50% and B shrank 20%, the average change is +15%. That tells you nothing about the actual combined rate. Don't mix those.
Mistake 3: Treating 0% as "no data." A 0% from 1,000 people is heavy evidence. A 0% from 2 people is noise. Weighting handles this if you keep the bases in.
Mistake 4: Rounding too early. Round at the end. If you round 75.4% to 75% and 39.6% to 40% before weighting, your final number drifts.
Mistake 5: Using the simple average because it's in Excel's AVERAGE function. Excel won't stop you. The function doesn't know your bases. You do.
Practical Tips
What actually works when you're staring at two percentages in real life?
- Always write the bases next to the %. Even on a sticky note. Forces your brain to see the weight.
- If you only have percentages, say "unweighted" out loud. That honesty alone prevents most disasters.
- Use this quick mental check: bigger group = bigger pull. If your answer isn't pulled toward the big group's number, redo it.
- For surveys, report both: the simple average and the weighted one, labeled. Let readers see the gap.
- In spreadsheets, build the weighted column yourself. Count_A = pct_A * base_A. Then sum counts / sum bases. Don't trust a one-cell AVERAGE.
- When presenting, show the base sizes in the footnote. People trust numbers more when they can see the weight behind them.
And look — if you're averaging two percentages from the same source with the same denominator (like two semesters of the same class size), relax. Simple average is your friend. The trouble starts the moment the denominators split.
FAQ
How do you average two percentages with different sample sizes?
You multiply each percentage by its corresponding sample size, add those two products together, then divide by the total sample size. This gives you a weighted average that reflects the actual volume behind each figure rather than treating both as equal.
Can I just take the midpoint if the sample sizes are close? Only if you're willing to be roughly right. "Close" is subjective — a base of 480 versus 520 is near enough that the simple average won't embarrass you, but 300 versus 700 is not. When in doubt, weight it. The extra line of math costs seconds and buys accuracy.
What if one percentage is based on a tiny sample? Then its pull on the average should be tiny — and weighting automatically handles that. A 100% from 3 responses shouldn't dominate a 62% from 900. If your weighted result still feels skewed by the small sample, report it separately and flag the low base rather than blending it into a headline number.
Why does my weighted average look lower than expected? Because the bigger group usually has the lower rate. That's not a bug; it's the whole point. Weighting reveals where the majority of the underlying instances actually sit, and if most of your volume came from the less impressive performance, the honest combined number will reflect that.
Is there a rule for when simple averaging is acceptable? Yes: identical or effectively identical bases, or when you are explicitly describing the average of the rates as rates (not the average of the underlying behavior). If you can't state the bases are equal, don't use the simple mean.
In the end, averaging two percentages is less about the arithmetic and more about respecting what sits underneath the numbers. That said, a percentage with no base attached is a story with no context — and blending two such stories without accounting for their size is how quiet errors turn into confident, expensive decisions. Weight when the bases differ, label what you did, and keep the denominators visible. Do that, and the math takes care of itself.
