Best A/B Testing Tools for Shopify (2026)
A neutral comparison of A/B testing tools for Shopify: native pricing apps, client-side platforms, and what checkout extensibility blocks.
📚 This article is part of the guide CRO Tools Compared: A Neutral Guide by Category (2026).
A Shopify store can A/B test almost anything on the home page, collection page, product page, and cart with a standard script tag. What it can no longer do, since Shopify’s move to checkout extensibility, is inject that same script inside the checkout itself: that area now only accepts Checkout UI Extensions, sandboxed components that Shopify builds, validates, and renders. This guide separates what each category of tool can actually test on a Shopify store in 2026: native pricing apps from the Shopify App Store, generic client-side platforms like VWO, Kameleoon, Convert, Optimizely, and Donnu, and exactly what checkout extensibility blocks for all of them equally. For a broader comparison of experimentation platforms outside the Shopify-specific context, see our neutral guide to CRO tools.
How A/B Testing Actually Gets Installed on a Shopify Store
There are three technical paths to running an A/B test on a Shopify store, and they are not interchangeable.
- Script tag or theme snippet. The most common path for generic A/B testing tools (VWO, Kameleoon, Convert, GrowthBook client-side, Donnu). A JavaScript snippet gets added to theme.liquid, or a lightweight app injects it for you. The script loads in the visitor browser, decides the variant, and swaps the DOM after the page has already started rendering, the same client-side model covered in our tool comparison guide.
- Native app from the Shopify App Store. Installs like any other store app, usually with API-level access to native Shopify objects (product, variant, price, cart). This removes setup work specifically for price and offer tests, because the app already understands Shopify’s data structure instead of relying only on generic CSS selectors on the rendered page.
- Checkout UI Extensions. The only path today to put anything inside the checkout pages themselves (Information, Shipping, Payment) or the Order Status / Thank You page. It works through standard extension points available on any plan, or deeper layout customization on Shopify Plus, built in TypeScript and React inside Shopify’s official sandbox. Very few generic A/B testing tools build this kind of extension, since the development model is entirely different from injecting a script.
What Changed in Shopify Checkout (and Why It Limits A/B Testing)
Shopify discontinued checkout.liquid in phases. According to Shopify’s official changelog and help center, the Information, Shipping, and Payment pages stopped accepting checkout.liquid customization on August 13, 2024, for every store regardless of plan. The Order Status / Thank You page, along with script tags and additional scripts on checkout, follows a separate, plan-dependent timeline: Shopify Plus stores had to complete the upgrade by August 28, 2025, while stores on a non-Plus plan have until August 26, 2026 before checkout.liquid, additional scripts, and script tags stop working there too. In place of checkout.liquid, Shopify offers checkout extensibility: a model where any checkout customization has to be built as a Checkout UI Extension, a sandboxed TypeScript and React component that Shopify controls, validates, and renders inside a secure environment, instead of arbitrary HTML, CSS, and JavaScript injected into one monolithic template.
In practice, this means two things for anyone testing a Shopify store in 2026:
- Outside checkout (home, collection, product, cart), nothing changed. A client-side script tag from any tool, Donnu included, still swaps copy, image, displayed price, banner, and layout normally, exactly as it would on any other website.
- Inside checkout, most generic A/B testing tools simply do not reach anymore. Building a Checkout UI Extension is a separate engineering project (TypeScript, React, review by Shopify itself), so few A/B testing vendors build and maintain one as part of their product. When an App Store listing advertises testing “in checkout,” it is worth confirming whether it actually uses an approved Checkout UI Extension, or is limited to testing what happens up to the cart, before the redirect into checkout.
This is not a limitation specific to Donnu or any single vendor: it is a platform-wide change from Shopify itself, documented officially, that affects every client-side A/B testing tool equally.
Native Pricing Apps vs Generic Client-Side Tools
Two different categories compete for this space, and picking the wrong one for your goal is the most common mistake.
Native pricing and offer apps (Shopify App Store)
The App Store has a dedicated collection of A/B testing and experimentation apps. Intelligems and Neat A/B Testing are among the more visible names, sharing a common focus: testing price, offers, shipping thresholds, and in some cases theme content, leveraging the fact that they already run inside the Shopify ecosystem and understand product, variant, and cart natively instead of rebuilding that logic with generic JavaScript. Intelligems, for example, publishes tiered plans (a lighter content and theme tier, with price, discount, and checkout testing gated to higher tiers) and reports profit-per-visitor accounting for cost of goods, discounts, and shipping, not just raw conversion. Neat A/B Testing takes a simpler approach: it rotates a product’s price, copy, image, or layout on a schedule (for example, every 24 hours) between the original and the test version and reports the funnel back on a dashboard. Since exact pricing tiers on both change over time, confirm the current plans directly on the App Store listing before comparing cost.
Generic client-side tools
VWO, Kameleoon, Convert, Optimizely, GrowthBook (self-hosted or with a cloud layer), and Donnu install through a script tag or theme snippet, the same mechanism they would use on any site that is not Shopify. VWO in particular ships a dedicated Shopify app that auto-deploys its tracking script across the store, explicitly excluding checkout, post-purchase, and order status pages, and streams native Shopify events (product views, cart actions, purchases) into its dashboard. The advantage of this category is breadth: any visual element of the store (home banner, offer popup, product page copy, collection ordering) can become a test variant without depending on a dedicated pricing app. The limitation is the mirror image of that advantage: these tools have no native understanding of the Shopify catalog, so testing an actual price change (not just displayed copy) usually requires extra integration work, and none of them reach the checkout after the move to checkout extensibility.
| Category | Installs as | Tests price/offers | Tests product/collection/home | Tests inside checkout |
|---|---|---|---|---|
| Native pricing apps (Intelligems, Neat A/B Testing, and similar) | Shopify App Store app | Yes, this is the core focus | Partial, depends on the app; some also cover theme and content | Varies by app and plan tier; confirm on the listing before assuming coverage |
| Generic client-side tools (VWO, Kameleoon, Convert, Optimizely, GrowthBook, Donnu) | Script tag or theme snippet | Partial, swaps the displayed price, not the underlying billing logic | Yes, any visual element | No, blocked since the move to checkout extensibility |
| Dedicated Checkout UI Extensions | Shopify-approved extension (in-house build or a specific app) | Not the purpose; typically used for complementary offers and fields | No, scope is checkout only | Yes, the only real path today |
Choosing by Criterion: Native Integration, Rigor, Pricing, and Setup
Instead of ranking vendors, match the tool to the criterion that decides your outcome. These four axes cover most of what differentiates the options above.
| Criterion | What to check |
|---|---|
| Native Shopify integration | Does the tool read product, variant, price, and cart objects directly, or does it only see the rendered page through CSS selectors? Native apps win here for price and offer tests specifically. |
| Statistical rigor | Does the tool publish its significance methodology (sample size guidance, peeking protection), or does it just show raw conversion counts per variant and leave interpretation to you? |
| Pricing model | Per-tested-visitor, flat monthly tiers gated by feature (as with Intelligems), or a broader platform subscription (as with VWO, Kameleoon, Convert, Optimizely)? Confirm current tiers directly with the vendor, published pricing in this category changes often. |
| Ease of setup | A script tag or a one-click App Store install versus a project that needs a developer, particularly true for anything that touches checkout, which now always requires a Checkout UI Extension regardless of which tool you pick. |
Weigh these before brand recognition. A tool that scores well on the criterion that matters most for your specific test (a price experiment versus a homepage banner test versus a checkout field) will outperform whichever name shows up first in a search, Donnu included.
Where to Focus: Product Page, Cart, and Checkout-Adjacent Tests
With the checkout area largely closed to generic tools, the highest-leverage testing work on a Shopify store concentrates on what happens before it, where the snippet still has full access:
- Product page. Title, hero image, photo order, social proof (reviews), add-to-cart button copy, free-shipping or low-stock badges. Usually the highest-leverage surface, because this is where the add-to-cart decision happens.
- Collection page. Product ordering, visible filters, items per row, presence of a discount or best-seller badge.
- Home banners and sections. Main hero, seasonal campaign banner, position of the most-searched category block.
- Offer and capture popups. Display timing, first-purchase discount offer, single-field versus full form.
- Cart (drawer or page), which is where most cart abandonment friction shows up. Free-shipping threshold messaging, complementary product upsell, stock-driven urgency copy. This is also the natural place to test recovery mechanics before a visitor ever reaches checkout, since anything inside the checkout pages themselves now requires a dedicated extension.
What is left out, or needs a dedicated extension to reach: anything inside the Information, Shipping, and Payment pages, and the Order Status / Thank You page. That includes tests many teams would like to run there, such as checkout field variations, last-minute offers, or confirmation copy; today those require building and getting a Checkout UI Extension approved, not installing a script. For a fuller process on prioritizing what to test first outside checkout, see our complete guide to conversion rate optimization.
How Much Traffic Does a Shopify Store Need Before Testing?
Regardless of which tool you pick, the question that decides whether a test is worth running is always the same: do you have enough visitors to detect the lift you want, in a reasonable amount of time? Enter your store’s current conversion rate and the improvement you want to detect below.
Two-proportion normal approximation, 2 variations (50/50). Tweak the inputs and watch it update live.
As a worked example: a product page with a 3% add-to-cart rate that wants to detect a 15% relative lift (moving to roughly 3.45%), at 95% confidence and 80% power, needs approximately 24,200 visitors per variant. At 1,000 visitors per day per variant (2,000 total daily visitors split evenly across two variants), that test finishes in about 25 days. The same test on a lower-traffic store, at 300 visitors per day per variant, takes about 81 days, close to three months, which is often the point where it makes more sense to target a bigger, bolder change (a larger MDE) than to chase a small lift that would take a full quarter to confirm. Run your own numbers in the calculator above before committing to a test plan; the relationship between the effect size you want to detect and the sample it requires is steep, not linear.
To see the same underlying math applied to declaring a winner once the test is running, not just sizing it beforehand, see our guide to statistical significance in A/B testing.
Common Mistakes When Testing on Shopify
| Mistake | Why It Backfires |
|---|---|
| Assuming an App Store “A/B testing” app covers checkout | Most generic apps and even some native pricing apps stop at the cart; confirm the exact scope on the listing before planning a checkout test around it |
| Testing price with a generic client-side tool without an integration layer | Client-side tools can swap the price shown on the page, but changing what a customer is actually billed usually needs native catalog access, which is what a pricing-specific app is built for |
| Running a test on traffic far below what the sample-size math requires | An underpowered test on a low-traffic store does not give a “less precise” answer, it gives a misleading one; target a bigger lift or accept a longer runtime instead |
| Comparing apps only by monthly price | Some native pricing apps gate price and checkout-adjacent testing behind a higher tier than the entry plan; the advertised starting price is not always the price that unlocks the feature you actually need |
Automate This on Donnu
You just saw the actual boundary generic A/B testing tools run into on Shopify: full access outside checkout, none inside it without a dedicated extension, and a real sample-size cost to testing on low traffic. If what you want to test lives on the product page, collection, home, cart, or a popup, and you do not have a developer on hand to build and maintain a custom integration for every test, a lightweight client-side tool with a no-code visual editor covers most of that work without turning into an engineering project.
Donnu is one option in that category, not the only one: it installs with a script tag on the theme like any generic tool in this guide, with a visual editor for swapping copy, images, and layout without code, Bayesian statistical reporting, and per-account data isolation from day one. If your case is specifically testing price and offers, a native pricing app from the App Store may serve you better. If your case is testing inside checkout itself, no generic tool, Donnu included, solves that today, that path runs through a dedicated Checkout UI Extension. For everything else, which is most of the buyer journey, start a free 14-day trial and compare it directly against whatever you use today.
Read also: CRO tools compared: a neutral guide | Best free A/B testing tools in 2026
Leia em português: Ferramentas de teste A/B para Shopify no Brasil.
References
- Shopify. Technical documentation on the checkout.liquid file and its deprecation. shopify.dev/docs/storefronts/themes/architecture/layouts/checkout-liquid.
- Shopify Changelog. “The checkout.liquid theme file is being deprecated,” with both deprecation phase dates. changelog.shopify.com/posts/the-checkout-liquid-theme-file-is-being-deprecated.
- Shopify Help Center. Upgrading and replacing your Thank You and Order Status pages, non-Plus deadline (August 26, 2026). help.shopify.com/en/manual/checkout-settings/customize-checkout-configurations/upgrade-thank-you-order-status/upgrade-guide.
- Shopify Help Center. Upgrading and replacing your Thank You and Order Status pages, Shopify Plus deadline (August 28, 2025). help.shopify.com/en/manual/checkout-settings/customize-checkout-configurations/upgrade-thank-you-order-status/plus-upgrade-guide.
- Shopify Partners Blog. “Checkout Extensibility Opens New Ways to Customize Checkouts on Shopify.” shopify.com/partners/blog/checkout-extensibility.
- Shopify App Store. A/B testing and experiments app collection. apps.shopify.com/collections/ab-testing-experiments.
- Shopify App Store. Intelligems: A/B Testing app listing. apps.shopify.com/intelligems.
- Shopify App Store. Neat A/B Testing app listing. apps.shopify.com/neat-ab.
- VWO. Shopify integration overview. vwo.com/integrations/shopify.
Frequently asked questions
- Can I run a generic A/B testing tool inside the Shopify checkout?
- Not anymore, in practice. Shopify retired checkout.liquid in phases: the Information, Shipping, and Payment pages stopped accepting theme customization on August 13, 2024, for every store regardless of plan. The Thank You and Order Status pages, including script tags and additional scripts, follow a plan-dependent deadline: Shopify Plus stores had to upgrade by August 28, 2025, while stores on a non-Plus plan have until August 26, 2026. Any checkout customization now has to be built as a Checkout UI Extension, a sandboxed component that Shopify itself validates and renders. Outside the checkout, on the home page, collection, product page, and cart, a generic client-side snippet still works exactly as before.
- What is the difference between a native Shopify pricing app and a generic tool like VWO or Donnu?
- A native app installed from the Shopify App Store, such as Intelligems or Neat A/B Testing, already understands your catalog, price objects, and cart natively, so it is usually faster to set up a price or offer test without writing code. A generic client-side tool installs through a script tag or theme snippet and can test almost any visual element of the store outside checkout (copy, images, layout, banners), but it has no native understanding of Shopify pricing logic and depends on its visual editor or custom JavaScript to touch price at all.
- How much traffic does a small Shopify store need to run a reliable A/B test?
- It depends on your current conversion rate and the size of the improvement you want to detect, there is no fixed number that applies to every store. As a reference point, a store with a 3% product-page add-to-cart rate that wants to detect a 15% relative lift needs roughly 24,200 visitors per variant at 95% confidence and 80% power, which is close to 25 days of testing at 1,000 daily visitors per variant. A store with meaningfully less traffic than that should target a bigger lift or accept a longer test; run your own numbers in the calculator further down this article.
- Is Google Optimize still an option for testing a Shopify store?
- No, Google Optimize was fully discontinued in September 2023 and is not available in any form. Since March 2026, Google Firebase A/B Testing stopped being mobile-app exclusive and extended to websites as well, with random assignment and built-in significance calculation, which makes it worth a look before committing to a dedicated platform, but it has no native integration with Shopify catalog or pricing data the way an App Store app does.
- Do Shopify A/B testing apps support statistically valid experiments, or just A/B toggles?
- It varies by app, and this is worth checking before you commit budget to one. Some native pricing apps publish their statistical methodology (sample size guidance, significance thresholds) directly in their documentation, while others simply report raw conversion numbers per variant and leave the interpretation to you. Generic client-side platforms built around statistics as their core product, frequentist or Bayesian, are more likely to guard against calling a winner too early. Confirm this directly in each vendor documentation rather than assuming any app that says "A/B testing" runs a properly powered test.