A complete experimentation platform
Split Test Pro brings together the variant types, statistical rigor, and AI tooling that the biggest ecommerce teams pay enterprise prices for — at a price that makes sense for any store.
CSS, JavaScript & redirect variants
Test anything from a button color to a complete page replacement. Style changes with CSS, content and DOM rewrites with JavaScript, or A/B two entirely different URLs with redirect variants.
AI-powered anomaly detection
Split Test Pro watches every running experiment for sample-ratio mismatches, traffic drops, novelty effects, and trend divergence — surfacing problems before they cost you a quarter.
AI pre-launch QA
Before you go live, our AI reviews your hypothesis, targeting rules, and metrics — flagging setup issues and estimating time-to-significance so you don't waste traffic on broken tests.
Bayesian statistics
Beta-Binomial models for conversion tests, Normal distributions for revenue and engagement. Get probability-to-be-best and credible intervals — not p-values you have to interpret.
Continuous metrics
Move beyond binary conversions. Track revenue per session, AOV, and engagement value — Split Test Pro models them with full distributions, not just averages.
AI results review
When an experiment ends, get a Claude-powered analysis with key findings, a clear recommendation (ship, iterate, or revert), and a projected revenue impact.
Variant screenshots
Auto-captured renders of every variant against your real page, so you can verify your CSS or JS actually applied — before a single visitor sees it.
Device & segment analysis
Slice results by desktop, tablet, and mobile to see how variants perform across audiences. Targeting, too: run tests for specific devices only.
Precise URL & event targeting
Target by full URL, host, path, or fragment — and defer activation until a click, scroll, or custom event fires. Test only the visitors you care about.
Custom traffic allocation
50/50 by default — or weight any way you want. Roll out risky variants to 10%, run multi-variant tests, or scale traffic to a winner gradually.
Works on any website
Drop a single script tag on any site to start testing — or install our Shopify app for a deeper integration with native funnel tracking. Same dashboard, same features, same stats either way.
Google Analytics 4 integration
Fires the standard experience_impression event so you can build GA4 audiences, segment your funnel by variant, and combine Split Test Pro with your full GA reporting.
Team workspaces & invites
Invite teammates with Owner or Member roles, manage multiple stores from one account, and collaborate on experiments without sharing logins.
Real-time results dashboard
Results stream in continuously as events arrive — no batch jobs, no waiting overnight. Watch probability-to-be-best update live.
Three variant types, one platform
Most A/B testing tools force you into a single mode — visual editor only, or hand-written JS only. Split Test Pro lets you mix and match: CSS for safe styling tweaks, JavaScript for content and DOM rewrites, and redirect variants when you want to compare two completely different pages.
Every variant is sandboxed and screenshot-verified before it goes live, so you catch broken selectors before your customers do.
AI that catches what you'd miss
Split Test Pro runs continuous anomaly checks on every live experiment. We detect sample-ratio mismatches, sudden traffic drops, zero-conversion variants, novelty effects, day-of-week patterns, and runaway tests — before they wreck your data.
Before launch, an AI review checks your hypothesis, targeting, and metrics for setup issues and estimates time-to-significance. After the test, it summarises the result and projects the revenue impact. You make the call — but you make it informed.
Bayesian stats, in plain English
Most A/B tools show p-values. Split Test Pro shows you what you actually want to know: "What's the probability this variant is the best?" and "What's the likely lift?" — backed by Beta-Binomial models for conversions and Normal distributions for revenue and engagement.
You see credible intervals, not significance thresholds. You see expected revenue impact, not asterisks. And the answer updates in real time as data streams in.
Live results
12% probability to be best · 95% credible interval [+24%, +66%]
88% probability to be best · 95% credible interval [+24%, +66%]
Up and running in minutes
Paste a script tag on any website, or install our Shopify app from the App Store — either way, you're collecting data the same day with no engineer required.
Same dashboard, same stats engine, same AI tooling, wherever your store lives. No platform lock-in.
Drop in anywhere:
<script src="https://cdn.splittestpro.app/stp.js" data-workspace="ws_abc123" async ></script>
Or install on Shopify in two clicks — same features, no code.
Ready to run your first test?
Sign up free in under a minute — no credit card, no installation steps until you're ready.