Conversion Rate Comparison Calculator
Put the visitors and conversions for two data sets side by side and get both rates plus the relative difference between them. The sanity check before you declare a winner.
Data set A
Data set B
About this tool
Each side is validated on its own, so you get a clear message if one set has zero visitors or more conversions than visitors rather than a silently wrong answer. When both sides are valid you see the two rates and the relative change of set B against set A, coloured green for an uplift and red for a decline.
If set A converts at exactly zero, the tool says so instead of printing a number. Dividing by a zero baseline gives you infinity, not a meaningful improvement, and that is a distinction worth keeping when the result is going into a decision.
Everything runs in your browser — nothing you enter here is sent to a server or stored. Provided as is, without warranty.
Reading a comparison without fooling yourself
The single most common misreading is treating relative and absolute change as interchangeable. If set A converts at 2% and set B at 3%, the absolute difference is one percentage point but the relative difference is 50%. Both are true. The 50% is the figure that ends up in the announcement, and it is also the figure that sounds enormous when the underlying move is a handful of extra signups. Get in the habit of stating both: “2% to 3%, a 50% relative lift” is honest, whereas “50% better” on its own is the kind of claim that quietly embarrasses you three months later.
This calculator tells you the size of a difference. It does not tell you whether the difference is real, and that is a separate question you have to answer yourself. Two variants with 5,000 visitors each and rates of 3.0% and 3.3% look like a clean win, but a swing of that size falls comfortably inside ordinary random variation at those volumes. Before acting, ask how many conversions each side actually produced, not how many visitors: under roughly 100 conversions per variant you are usually still looking at noise, and a proper significance test is the right next step.
The less obvious trap is the peeking problem. If you keep checking a running test and stop the moment the numbers look good, you will declare winners that are not there, because at some point during almost any test the two arms drift apart by chance. Decide the sample size and the end date before you start, and only then open the results. Related: never compare two periods that were exposed to different conditions. Last week against this week is not a fair fight if a newsletter went out, a public holiday landed, or a paid campaign changed the mix of who arrived. Split the traffic concurrently or accept that you are looking at a rough indication.
Finally, check that both sides are counted the same way. It is surprisingly easy to end up with one variant measured on sessions and the other on users, or with bot traffic filtered out of one report and not the other. If you need to work out either rate first, the conversion rate calculator handles the single-set case, and the times increase calculator converts the gap into a multiplier if that framing suits your audience better.
Ready to see where agents can take cost out of your business?
Tell us about the process you want to optimise. Vlad personally reviews every brief and replies within one business day.