Skip to main content

Helpfulness score calculator for "Was this helpful?" answers

Free tool. A helpfulness score is the share of people who answered Yes to "Was this helpful?": Yes answers divided by all answers. Enter your counts to get the score, the range it is likely in, and whether you have enough answers to judge.

  • Free
  • No sign-up
  • Runs in your browser
What do you want to check?

Your helpfulness score

75% 45 of 60 answers said Yes.

  • The true score is likely between 63% and 84% (95% range).
  • Enough answers to judge this page.

Ask "Was this helpful?" on every page

Asklet adds a short question to the end of each article or docs page and shows the answers for each page, so you can see which pages help and which need work.

How to use the helpfulness score calculator

  1. Enter your Yes and No answers

    Take the counts for one page and one period from your help centre, docs tool or feedback widget.

  2. Read the score and its range

    The range shows where the true score is likely to be. Below 30 answers, read the score as a hint.

  3. Compare two pages or periods

    Switch to Compare two to see whether a gap, such as before and after a rewrite, is bigger than chance.

How to calculate a helpfulness score

Divide the Yes answers by all answers and multiply by 100. A page with 45 Yes and 15 No answers scores 45 ÷ 60 = 75%. Geckoboard's definition of the help centre metric "percentage of positive votes" is the same sum.

The score describes the people who answered, not everyone who read the page. Most readers never answer, and the ones who do may not be typical of the rest, so compare pages with each other rather than reading a score as the opinion of all your readers.

Why the range matters more than the score

Three Yes answers and one No make 75%, and so do 300 Yes and 100 No. The first could really be anywhere from 30% to 95%; the second is between 71% and 79%. The calculator shows that range, so a handful of early answers does not look like a verdict.

The range is a 95% Wilson score interval, first described by Edwin Wilson in 1927. Brown, Cai and DasGupta compared the usual ways of putting a range on a percentage and recommend the Wilson interval (or the Jeffreys interval) for small samples. The plus-or-minus formula found in most textbooks behaves erratically there, and they found the textbook rules for when it is safe cannot be trusted.

Below 30 answers, the calculator still shows the score but calls it a hint, not a verdict. That floor is our rule of thumb, not a statistical law: what matters is whether the range is narrow enough for the decision you want to make.

What is a good helpfulness score?

There is no published benchmark we could verify. A score depends on the kind of page and on how the question is worded, so a number from another site tells you little about yours. Set your own baseline instead:

  • Work out the score of every page that has enough answers
  • Take the middle value (the median) as your baseline
  • Look first at pages well below it that many people read
  • Check whether a page moves after you change it, with Compare two

Compare like with like: the same question wording and the same dates on every page.

Which pages to fix first

A low score on a page few people read matters less than a middling score on a page thousands read. Docsy's guide to "Was this page helpful?" feedback says to start with the pages that get a lot of page views and more No than Yes answers.

To rank pages, multiply each page's share of No answers by its page views for the same dates. If the people who answered are typical, that is a rough count of readers the page did not help. Our docs feedback triage guide adds how much each page matters and what the fix costs.

Comparing two pages or two periods

To check whether a rewrite worked, compare the page before and after. The calculator works out the difference and its 95% range with Newcombe's method, which combines the two Wilson intervals. In his 1998 comparison of eleven methods, it performed well and can be used whatever the number of answers.

If the range includes zero, the gap could be chance. 18 Yes out of 24 against 14 out of 24 is a 17-point gap that could be chance. 180 out of 240 against 140 out of 240 is the same gap, and it is bigger than chance.

Common questions

What is a helpfulness score? The share of people who answered Yes to a "Was this helpful?" question on a page: Yes answers divided by all answers, as a percentage. Help centre tools also call it the helpfulness rate or the percentage of positive votes.

How do you calculate the helpfulness percentage? Divide the Yes answers by the Yes and No answers together, then multiply by 100. 45 Yes and 15 No answers give 75%.

What is a good helpfulness score? There is no published benchmark we could verify, and scores depend on the page and the question. Use the median of your own pages as a baseline, and start with the pages well below it that many people read.

How many answers do you need? Enough for the range to be narrow enough for your decision. As a rule of thumb, wait for 30 answers before you judge a page: with 4 answers, a 75% score could really be anywhere from 30% to 95%.

Why does my helpfulness score jump around? Because each answer moves a small total a lot. With 4 answers, one more No takes a page from 75% to 60%. The range narrows as answers come in.

How does Asklet show helpfulness? Asklet groups answers by page. Counts show from the first answer, and percentages wait until a page has enough answers to be meaningful, so a handful of early answers don't look like a trend.

Do you store what I enter? No. The calculator runs in your browser and nothing is sent to us.

en