Click-Through Rate (CTR) Calculator
CTR with its confidence range, and the impressions it would take to tell your rate apart from another one, because small rates are noisier than they look.
Impressions 142,000
Clicks 720
Click-through rate 0.507%
95% range 0.471% to 0.545%
Impressions per click 197
Spend $1,022.00
Cost per click $1.42
Cost per thousand impressions $7.20
Rate to compare with 0.55%
Difference -0.043 points
Relative difference -7.8%
Impressions each side to be sure 447,222
You have 142,000
720 clicks from 142,000 impressions is 0.507%, and the 95 percent range
on this many impressions is 0.471% to 0.545%. A margin of a tenth of a
percentage point at this rate takes about 19,380 impressions, and you
have more, which is why the range is this tight. Halving it again takes
four times what you have, or 568,000 impressions.
Telling 0.507% apart from 0.55% takes about 447,222 impressions on each
side. You have 142,000, which is not enough: the -7.8% relative
difference may be entirely noise, and rebuilding the creative around it
would be building on a coin flip.
The three ad numbers are one identity: a $1.42 cost per click at 0.507%
is a $7.20 cost per thousand impressions. A rising CTR at a flat CPM
shows up as a falling CPC, which is the same event described three ways.
A CTR is only comparable within a placement. Search, display, email and
a link in a sidebar differ by more than an order of magnitude, and so do
a rate measured against served impressions and one measured against
viewable ones. Compare the same ad slot over time, not two slots against
each other.
Output is valid and updates as you type.
Fix the highlighted fields to update the output.
Clicks divided by impressions. The reason this page has a second half is that click-through rates are small, and small rates are noisier than they look.
Two ads at 0.51 and 0.55 percent on 142,000 impressions each look like an eight percent difference. They are indistinguishable: telling those two rates apart takes around 450,000 impressions on each side. Calling the winner at 142,000 and rebuilding the creative around it is a quarter spent on a coin flip, so this prints the interval and the impressions the comparison would actually need.
How to use
- Put in the impressions and the clicks.
- Add the spend for the cost per click and the CPM.
- Put another ad’s rate in the comparison field to see whether the difference between them is real yet.
Example
720 clicks from 142,000 impressions, 1,022 of spend, compared against another ad at 0.55 percent:
Impressions 142,000
Clicks 720
Click-through rate 0.507%
95% range 0.471% to 0.545%
Impressions per click 197
Spend $1,022.00
Cost per click $1.42
Cost per thousand impressions $7.20
Rate to compare with 0.55%
Difference -0.043 points
Relative difference -7.8%
Impressions each side to be sure 447,222
You have 142,000The relative difference reads as 7.8 percent worse, and the range on this ad alone already reaches 0.545. Nothing has been proven. Three times the impressions would settle it.
Pitfalls
Relative differences on small rates are theatre. Going from 0.5 to 0.6 percent is “a 20 percent improvement” and one extra click per thousand impressions. Both descriptions are true; only one of them is worth a headline.
CTR is only comparable within a placement. Search, display, email, a newsletter sponsorship and a sidebar link differ by an order of magnitude or more. So does a rate measured against served impressions versus viewable ones. Compare one slot to itself over time.
A rising CTR is not automatically a win. Clickbait creative raises the rate and lowers everything after it. Look at the cost per conversion in the same breath, or you will optimise for people who click and leave.
Position and audience move the rate more than creative does. The same ad in the first search result and the fourth will differ by more than most creative tests produce, which is why a test that changed both at once has told you nothing.
Bots and accidental clicks inflate it. Mobile display in particular carries mis-taps. If a channel has a high CTR and a terrible conversion rate, suspect the clicks before you blame the landing page.
Tiny counts make percentages look precise. Two clicks from 300 impressions is 0.667 percent to three decimal places, and it is also consistent with anything from 0.1 to 2.4 percent.
Compatibility
Arithmetic in the browser: nothing is uploaded and nothing is stored. The share link carries the figures, which is the fastest way to answer “is this difference real?” in a thread.
The range is a Wilson score interval. The textbook p ± z·√(p(1−p)/n) fails on exactly this kind of data: at small proportions it is too narrow, and it returns negative lower bounds. Wilson’s stays inside 0 and 100 percent at 720 clicks and at 7.
The “impressions each side to be sure” figure is the standard two-proportion sample size at 95 percent confidence and 80 percent power. It is a planning number: trust the order of magnitude, not the second digit. It also assumes the two ads are running against comparable traffic, which a test with one ad on a weekend and the other on a Monday is not.