Frequently asked questions

No. SplitPea works fine with smaller websites, as long as you treat testing as a way to learn from real visitors rather than a magic button for instant certainty.

Lower traffic usually means you should test one meaningful change at a time, let experiments run long enough to collect a fair sample, and be comfortable with "not enough data yet" as a real answer. SplitPea keeps those states visible so you are not pushed into calling a winner too early.

You can test headlines, button copy, signup forms, pricing page wording, landing page layouts, call-to-action sections, lead forms, and other page elements where you are not sure which version will work better.

The best tests are tied to a clear visitor action: sign up, book a call, submit a form, click a pricing button, or start checkout. If a change could plausibly affect that action, it is usually a candidate for an experiment.

If you want a starting list, read our guide to what to A/B test on your website.

No. Think of SplitPea as an experimentation layer. Analytics tools tell you what happened across your site. SplitPea helps you test a specific change and decide whether that change helped.

You can keep your normal analytics setup for traffic sources, page views, and broader reporting. SplitPea focuses on the narrower workflow: create a variant, split visitors, track the goal, read the result, and keep a record of the decision.

Yes. SplitPea is useful for freelancers and agencies who want to test campaign pages, service pages, pricing pages, and lead-generation pages across client sites.

The main benefit is having a clear answer when a client asks whether a change worked. Instead of relying on taste or preference, you can show the experiment, the variants, the conversion goal, and the outcome history.

No. You need to be able to add the SplitPea snippet to your site once, or ask someone who manages the site to add it for you. After that, the visual editor lets you select page elements and create simple variants without writing code.

If you are comfortable with CSS selectors, you can use them for more precise targeting. If not, the point-and-click flow is designed for normal website operators, marketers, founders, and client teams.

SplitPea uses Bayesian statistics to estimate how likely it is that one variant is outperforming another. In the interface, that is shown as confidence, alongside visitors, conversions, conversion rate, and lift.

The important part is the decision state. SplitPea helps distinguish a likely winner from a test that is still running, inconclusive, or too light on data. That keeps the result useful without asking you to become a statistician.

Our plain-English confidence guide explains the 90% threshold in more detail.

It depends on your traffic and how big a difference you're testing for. Some tests wrap up in a few days. Others need a couple of weeks.

SplitPea shows you progress as data comes in, so you're never just staring at a blank screen wondering. And if a test is going to need more time, it'll say so.

That's a real result. It means the change you tested didn't make a meaningful difference, which is useful to know.

SplitPea will tell you when a test is inconclusive so you can stop it and move on to testing something else, instead of waiting around hoping the numbers shift.

The snippet is under 8KB and loads asynchronously. Your visitors won't notice it.

Yes. There are no contracts and no cancellation fees. You can cancel from your account settings whenever you want.

Yes. The Starter plan is free and doesn't require a card. You can run one experiment on one site with up to 500 visitors per month, no strings attached.

Your experiments keep running, but we'll stop counting new visitors for that billing cycle. No surprise charges. If you're consistently hitting the limit, it's probably time to move up a plan.

Yes. You can upgrade or downgrade at any time. If you upgrade mid-cycle, we'll prorate the difference. If you downgrade, the change takes effect at your next billing date.

A visitor is a unique browser that sees one of your active experiments during a billing cycle. If the same person visits twice, that's one visitor. People who visit pages without an active experiment don't count.

No. All plans are month-to-month. Cancel whenever you want and you won't be billed again.

Yes. Every plan gets the visual editor, Bayesian confidence scoring, goal tracking, decision history, and email support. The only differences are how many sites and experiments you can run and your visitor limit.