Free Shopify split testing: every $0 path in 2026

Free split testing on Shopify is real, and in 2026 there are more zero-dollar paths than most merchants expect. The honest complication is that free of app fees and statistically valid are two separate things, and every free path hands the second one to you.

This guide maps every way to run a split test on Shopify that costs $0 in add-on fees: the native Rollouts experiment, three genuinely free apps, and the DIY duplicate-theme method. It is the free-tier companion to our full Shopify A/B testing guide, narrowed to the paths that cost nothing. It also covers the part vendors skip, which is what free leaves you to do yourself. Prices and ratings are verified July 2026 on the Shopify App Store and shopify.com/pricing; testing tools change tiers often, so re-check before you commit.

The five $0 paths, at a glance

TL;DR Five paths cost nothing in add-on fees: native Rollouts, Trident AB's free tier, Essential CRO Split AB Testing, Instant's free app, and the DIY duplicate-theme split. Not one ships a significance engine.

Here is the whole free landscape in one table, then a plain-English walk through each option and where it breaks.

Path	Cost (July 2026)	Significance engine	Best for
Shopify Rollouts	Free add-on; A/B needs Grow (~$105/mo, verify live)	None	Native server-side theme splits
Trident AB (free tier)	Free: 500 impressions/mo, 1 test	None	Learning the mechanics
Essential CRO Split AB Testing	100% free, no paid tier	None documented	Page-template splits, Built for Shopify
Instant A/B Testing	Free app (lightly reviewed)	None	Landing & product-page tests
DIY duplicate-theme	$0 in app fees; real dev time	None (you build it)	Full control, decays in production
StorePilot	Free first 3 months, then ~$129/mo (pre-launch)	Deterministic honest stats	Low traffic, done-for-you

Prices and ratings verified July 2026 on the Shopify App Store and shopify.com/pricing; they change often, so confirm before you commit. StorePilot is ours and pre-launch, disclosed as such. For the full paid field, see our 8-tool comparison.

Two patterns jump out. First, every free path leaves the statistics to you: no significance test, no automatic winner. Second, the native option is the only one Shopify runs on its own servers, which is what makes it the fastest and the safest, and also the one with a plan gate. That is where to start.

Shopify Rollouts: native, and truly free of app fees

TL;DR Rollouts splits theme versions on Shopify's servers with no add-on fee and no flicker, generally available since the Spring '26 Edition. The catch: the A/B experiment needs the Grow plan (about $105/mo, verify live), and there is no significance engine.

Rollouts is Shopify's native tool for scheduling, gradually publishing, and A/B testing theme changes. Because the split happens on Shopify's servers before the page is sent, the assigned version is there at first paint: no app to install, no flash of the old theme, and no drag on your Core Web Vitals. It became generally available with the Spring '26 Edition on June 17, 2026, and it lives under Markets > Rollouts, also reachable from Online Store > Themes (Shopify Help Center).

The honest catch is the plan gate. Scheduling and publishing theme changes work on Basic and up, but the actual traffic-split experiment, with its analytics, is gated to the Grow plan or higher, Shopify's second tier at about $105/mo US on monthly billing (verify live on shopify.com/pricing). After Shopify's 2026 rename the ladder reads Basic, Grow, Advanced, Plus, where "Grow" is the old middle "Shopify" tier. Older guides that say you need "Advanced" are stale; market-specific targeting is the piece that needs Advanced or Plus.

Rollouts is a splitter, not a judge. Its analytics show conversion rate, average order value, gross sales, sessions, bounce rate, and reached-checkout and add-to-cart rates. Four things are confirmed missing: a statistical-significance test, an automatic winner call, revenue per visitor, and any projected-dollar range (Shopify Help Center). Shopify's own changelog verb is "find the winner," which means you find it. The scope has hard edges too: it cannot test prices or discounts, because those are not theme changes; the unit of a test is a whole theme version, not one isolated section; and segmentation is market-level only. One footnote worth knowing: an Online Store theme experiment launched immediately collects no analytics, so you have to schedule it for the numbers to record. For the full walkthrough, see our deep dive on Shopify Rollouts, and if pricing is what you actually want to test, our Shopify price testing guide covers what Rollouts cannot reach.

Free A/B testing apps: Trident, Essential CRO, and Instant

TL;DR Three App Store apps run a split test at $0: Essential CRO Split AB Testing (100% free, Built for Shopify, 5.0 stars / 25 reviews), Trident AB's free tier (500 impressions a month, one test), and Instant's free app (lightly reviewed). None documents a significance engine.

Essential CRO Split AB Testing is the most credible genuinely-free option besides the native one. Built by Essential Apps (Vilnius, Lithuania) and launched June 11, 2025, it is 100% free with no paid tier, carries a Built for Shopify badge, and rates 5.0 stars across 25 reviews (Shopify App Store, verified July 2026). It runs client-side and splits page templates such as product, collection, and homepage, with a zero-flicker claim. It documents no significance engine and no published test caps, so read it as a clean, free way to run a layout split, not a verdict machine.

Trident AB has a real free tier: up to 500 impressions a month, a single A/B test, basic support, and test analytics. Its paid ladder runs Premium at $19.99/mo (5,000 impressions a month, unlimited simultaneous tests, 5-day trial), Business at $24.99/mo (50,000 impressions a month), and Enterprise at $49.99/mo (unlimited impressions), with annual billing saving 20% (Shopify App Store, verified July 2026). It rates 4.0 stars across 24 reviews. The free cap is the story: 500 impressions a month cannot reach significance on a real conversion gap, so the free plan is for learning the mechanics.

Instant (instant.so, Amsterdam) also lists a free A/B testing app that tests any landing or product page plus custom URLs, with in-app conversion tracking for sessions, add-to-cart, reached checkout, and completed checkout, and no significance engine. It is lightly reviewed, and its free-tier caps were not confirmable this week, so verify the current listing before you lean on it. For the full field including the serious paid tools, see our 8-tool comparison.

The DIY duplicate-theme split (free, if you have dev time)

TL;DR Duplicate the published theme, assign visitors by cookie, route a share to the variant, and tag orders by arm. It costs $0 in app fees but causes flicker, needs SEO safeguards you enforce yourself, and decays in production.

A Shopify store has exactly one published theme, so the DIY method works around that: you duplicate the published theme, edit the copy, assign each visitor to A or B with a cookie, route a share of traffic to the variant, and tag orders by arm so you can compare them later. It is genuinely free of app fees, and it gives you total control over what changes. The problems are all in production. Because the swap happens client-side, visitors see the live theme flash before the variant loads, and the variant's first load is slower.

Then there is SEO, which you have to police by hand. Google's website-testing guidance is specific: never cloak, set rel=canonical from the variant back to the original, use 302 (not 301) redirects, and remove the test promptly once it is done (Google Search Central). Rollouts and serious apps comply by design; your homemade script does not, unless you make it. There is also a quiet caching hazard: CDNs cache pages, not cookie assignments, so a visitor assigned to B can be served a cached A page and the split degrades without any error to warn you. QA fresh sessions across devices before you trust a single number. The blunt verdict is that the DIY split works in a demo and decays in production, so budget real developer time or skip it. Our Shopify theme A/B testing guide walks the full duplicate-theme method step by step.

What died: Google Optimize

TL;DR Google sunset Google Optimize on September 30, 2023, with no replacement. For theme-level tests, Shopify Rollouts is the closest free successor.

For years, the default free A/B testing tool for a lot of Shopify stores was not a Shopify tool at all: it was Google Optimize, bolted on through Google Tag Manager. Google sunset Google Optimize on September 30, 2023, with no replacement (Google Analytics Help), which pulled the free rug out from under everyone who relied on it. If you are searching for a free split tester today because your old setup stopped working, that is why.

For theme-level tests, Rollouts is the closest free successor: native, server-side, no add-on fee, and none of the tag-manager fragility that made Optimize flicker. If you are migrating off Optimize specifically, our Google Optimize alternative for Shopify guide maps the options side by side.

The honest cost of free

TL;DR No significance engine means you judge the result, and low traffic makes most free tests underpowered. The discipline is what makes free valid.

Free tools remove the invoice, not the two hard parts of testing. The first is sample size. A rough rule is that you need about 16 divided by (baseline conversion rate multiplied by relative lift squared) visitors per variant. At a 2% baseline conversion rate, detecting a 10% relative lift takes about 78,000 visitors per variant (formula from our A/B testing guide). Set that against Trident AB's free cap of 500 impressions a month and the mismatch is the whole story: the free plan cannot physically gather enough data to call a real conversion gap. If your store does a few thousand sessions a month, read the low-traffic playbook and the World Cup sample-size explainer for the intuition on why most small tests are underpowered before they begin.

The second hard part is discipline, and it gets sharper when the tool has no significance engine to hold the line for you. Without one, the temptation is to watch the dashboard and call a winner the moment B pulls ahead. That is peeking, and it is expensive: repeated peeking can push a real 5% false-positive rate to about 26% (Evan Miller). The fix is free and unglamorous: pick a sample size before you start, refuse to look for a decision until you reach it, and read the result once. Our guide on statistical significance covers how to do that without fooling yourself.

Low-traffic stores feel both problems at once, which is exactly the gap StorePilot (ours, pre-launch, said plainly) was built for. It watches real behavior, finds the friction worth testing, builds the variant, and judges it on revenue per visitor with deterministic honest stats: a minimum sample and a real significance bar, no early winners, on any plan. It is free for the first 3 months and then about $129/mo at the founding rate, and none of this guide depends on you using it, the free paths above are all real and shipping today.

The honest takeaway: free split testing on Shopify is genuinely useful, and often enough to catch a big, obvious win. What free never hands you is the judgment. Pick a sample size before you start, refuse to peek, and free testing stays honest.

Questions merchants keep asking

Can you A/B test on Shopify for free?

Yes, in several ways. Shopify Rollouts is native with no add-on fee, though the traffic-split experiment is gated to the Grow plan or higher (about $105/mo US, verify live on shopify.com/pricing). Trident AB has a free tier (up to 500 impressions a month, one test), Essential CRO Split AB Testing is 100% free, and the DIY duplicate-theme method costs nothing in app fees. None of them includes a statistical-significance engine, so you judge the result yourself (Shopify Help Center; Shopify App Store, verified July 2026).

Is Shopify Rollouts really free?

There is no add-on fee, and scheduling and publishing theme changes work on Basic and up. The catch is that the actual traffic-split experiment, with its analytics, is gated to the Grow plan or higher, Shopify's second tier at about $105/mo US on monthly billing (verify live on shopify.com/pricing). It became generally available with the Spring '26 Edition on June 17, 2026 (Shopify Help Center).

Can I split test on Shopify without an app?

Yes, two ways. Rollouts is native, so there is no app to install: find it under Markets > Rollouts (also reachable from Online Store > Themes). Or you can run the DIY duplicate-theme method, where you duplicate the published theme, assign each visitor by cookie, route a share to the variant, and tag orders by arm. The DIY route causes flicker and needs SEO safeguards you enforce yourself (Google Search Central website-testing guidance).

What is the best free A/B testing app for Shopify?

Among genuinely free apps on the Shopify App Store (verified July 2026): Essential CRO Split AB Testing is 100% free with a Built for Shopify badge and a 5.0-star rating across 25 reviews; Trident AB has a free tier (500 impressions a month, one test) and rates 4.0 stars across 24 reviews; Instant also lists a free A/B app but is lightly reviewed. None of the three documents a significance engine, so treat them as tools to learn the mechanics.

How many visitors do I need for a free split test to be valid?

Roughly 16 divided by (baseline conversion rate multiplied by relative lift squared) per variant. At a 2% baseline conversion rate, detecting a 10% relative lift takes about 78,000 visitors per variant (formula from our Shopify A/B testing guide). That is why Trident AB's free cap of 500 impressions a month cannot reach significance on a real conversion gap: free plans teach the mechanics, not the verdict.

Does free Shopify split testing include statistical significance?

No. Shopify Rollouts shows conversion rate, average order value, gross sales, sessions, bounce rate, and reached-checkout and add-to-cart rates, but it documents no significance test, no winner call, and no revenue per visitor (Shopify Help Center). The free apps document none either. Every free path hands the significance judgment to you, which is the discipline that decides whether a free test is trustworthy.

Is the DIY duplicate-theme method safe for SEO?

Only if you enforce Google's website-testing rules yourself: never cloak, set rel=canonical from the variant back to the original, use 302 (not 301) redirects, and remove the test promptly once it is done (Google Search Central). It also causes a visible flicker, and CDNs can serve a cached page to the wrong variant, which degrades the split silently. QA fresh sessions across devices before you trust the numbers.

Wasn't Google Optimize the free option?

It was, for years. Google sunset Google Optimize on September 30, 2023, with no replacement (Google Analytics Help). That left a gap for merchants who relied on it. For theme-level tests, Shopify Rollouts is the closest free successor today, since it splits live traffic natively with no add-on fee.

Can I test prices for free on Shopify?

No. Rollouts cannot test prices or discounts, because the unit of a Rollouts test is a whole theme version and prices are not theme changes (Shopify Help Center). The DIY duplicate-theme method does not touch pricing either. Price and offer testing needs a specialist such as Intelligems or ABConvert, which work around Shopify's price-test limits with cart-transform functions.

Is free split testing enough, or should I pay?

Free is a fine way to start and to catch a big, obvious win. The catch is two-fold: most free tests are underpowered at low traffic, and without a significance engine you carry the discipline yourself, including not peeking. Repeated peeking can push a real 5% false-positive rate to about 26% (Evan Miller). Pay for a significance engine once your traffic and the stakes of the decision justify it.