How much traffic do you need for A/B testing?

It depends on your conversion rate and the size of lift you want to detect, not on a single magic number. As a reference point: at a 2% conversion rate, detecting a 20% relative lift at 95% confidence and 80% power needs about 21,000 visitors per variant. At a 10% conversion rate the same test needs about 3,800 per variant. Sites with roughly 10,000 or more monthly visitors on the tested page can usually run meaningful tests on bold changes; below about 1,000 monthly visitors, classic A/B testing rarely concludes.

Can I A/B test with less than 1,000 visitors a month?

Rarely with classic significance testing. At that traffic, even an 8-week test can usually only detect lifts of 50% or more, and real winning variants rarely move a metric that much. Better options: test one bold redesign at a time and judge it over a longer window, pool several changes into one variant, test on your highest-traffic page instead, or use qualitative methods until traffic grows.

What should I do if I do not have enough traffic to A/B test?

Four practical moves: test bigger swings (a new offer, layout, or headline moves numbers far more than a button color, and big lifts need less traffic to detect), test where the traffic is (your highest-traffic page, not your favorite page), bundle several related changes into a single variant so there is one bigger effect to measure, and lengthen the test window while accepting that only large effects will conclude.

What is the minimum traffic for A/B testing?

There is no universal minimum, but useful rules of thumb: below about 10,000 monthly visitors on the tested page, only very large changes (roughly 30% lifts and up) are detectable, so classic A/B testing is hit and miss. Between 10,000 and 100,000 monthly visitors you can test bold changes such as layouts, offers, and headline angles. Above about 100,000 monthly visitors you can detect the smaller lifts that iterative testing produces. Enter your own numbers above rather than relying on the rule of thumb, because your conversion rate moves the threshold as much as your traffic does.

How do I calculate the minimum detectable effect (MDE) for my traffic?

The minimum detectable effect is the smallest relative lift a test can reliably detect given its sample size. This checker calculates it for you: it converts your monthly visitors into visitors per variant over 2, 4, 6 and 8 week windows, then solves the standard sample size formula backwards for the smallest MDE each window supports at 95% confidence and 80% power. If the MDE it reports is larger than the lift you realistically expect, the test cannot conclude and you need more traffic, a bolder change, or a longer window.

Does a higher conversion rate reduce the traffic I need?

Yes, substantially. The visitors needed to detect the same relative lift drop roughly in proportion to your conversion rate: a page converting at 10% needs about one fifth of the traffic that a 2% page needs for the same test. This is why testing a high-converting step (like an add-to-cart or signup step) is often feasible when testing a low-converting landing page is not.

Do You Have Enough Traffic to A/B Test? Free Checker

Monthly Visitors

Visitors per month on the page you want to test

Conversion Rate (CR%)

Current conversion rate of that page

Number of Variations

Including the control

Don't know your conversion rate? Start with an industry benchmark:

Your traffic verdict

Yes, but only for bold changes. A 4-week test at your traffic can only detect a lift of 44.8% or bigger. Test big swings like offers, layouts, and headlines. Skip subtle tweaks; they will never reach significance here.

For scale: detecting a typical 20% lift needs about 21,116 visitors per variant. At your traffic that is roughly 18 weeks of testing.

Test length	Visitors per variant	Smallest detectable lift
2 weeks	2,333	66%
4 weeks	4,666	44.8%
6 weeks	7,000	35.9%
8 weeks	9,333	30.8%

What kind of change wins this big?

Median winning lift by change type, from real tests on Mida: hero image 20.2%, styling 22.2%, body copy 22.8%, CTA copy 24.7%, layout 25%, price framing 27.5%, headline 29.1%, split URL 32.3%, social proof 32.5%, form 71.1%. Read this against the table above: when your smallest detectable lift is higher than a change type's median, most winners of that type would never reach significance on your traffic. Full data in the A/B testing benchmarks.

Mida plans this for you on every test and tells you when a result is real. Free for up to 100,000 tested users a month, which covers most sites on this page.

Try Mida free

Do You Have Enough Traffic to A/B Test?

How Much Traffic Do You Actually Need?

How To Read Your Verdict

What To Do When Traffic Is Not Enough

Try Our Other Free A/B Testing Tools

Frequently Asked Questions

The A/B testing platform for people who
care about website performance