Intelligems vs Shoplift: Which Shopify A/B Testing Tool, and How Much Traffic Do You Need?

What is the difference between Intelligems and Shoplift?

Intelligems is built around testing the economics of the store (prices, discounts, shipping, offers), and Shoplift is built around testing the storefront (themes, templates, pages). Both now overlap, but their roots still show. Intelligems describes Shoplift as "a visual, no-code solution that integrates into Shopify's Theme Customizer" and itself as a platform that "specializes in profit optimization" (Intelligems). Shoplift describes Intelligems as "a tool to let you run complex pricing experiments" and positions itself for "high-velocity experimentation" (Shoplift). Read both as vendor positioning.

Which tool should a supplement brand choose?

For most supplement brands running Meta, the highest-value tests are offer tests: subscription discount, bundle tiers, free gifts, shipping thresholds and price. That favors Intelligems. If your roadmap is mostly page structure (new product page template, landing page redesign, navigation), Shoplift is simpler to run.

• Choose Intelligems if you want to test subscribe-and-save discounts, quantity break pricing, free shipping thresholds or gift-with-purchase offers.
• Choose Shoplift if your team works in the theme customizer and wants to compare templates and landing pages quickly without a developer.
• Use Shopify Rollouts for occasional theme experiments if you do not need price or offer testing. Shopify's Rollouts can schedule theme, checkout and catalog changes and run an experiment that "compares a treatment against a control".

One caution with landing page tests: when the variable is the destination page for an ad, split at the ad level in Meta rather than inside a site tool. A site-level split only sees people after they land. See landing page vs product page for Meta ads.

How much traffic do you need for a Shopify A/B test?

Far more than most stores expect. The required sample depends on your baseline conversion rate and the smallest lift you care about detecting. The table below uses the standard two-proportion formula at 95% confidence (two-sided) and 80% power, which matches the figures published in testing roundups: Humblytics cites roughly 68,000 sessions per variant for a 10% lift on a 2.4% baseline (Humblytics), and Fudge cites about 50,000 monthly sessions on the tested page for a 10% lift on a 2% baseline to resolve "in a sensible window" (Fudge).

Double the per-variant number for a standard two-arm test. So a product page converting at 2% needs about 161,000 sessions to reliably detect a 10% lift, and about 42,000 to detect a 20% lift. At 20,000 sessions a month on that page, the 10% test would take eight months. That is why small-lift tests on small stores rarely produce real answers.

What should you test at your traffic level?

Match the size of the change to the traffic you have. Small traffic can only detect big effects, so test things that could plausibly move results by 20% or more.

A note on sequential changes for small stores: if you ship a change without a split test, compare the same weekdays before and after, keep ad spend and creative as stable as you can, and write down the date. It is weaker evidence than a test, but far better than guessing, and it lets you undo a change quickly if revenue per session falls.

Humblytics makes the same point for lower-traffic stores: test changes with larger expected effects (price, offers, shipping) or move testing upstream to ad creative and landing pages where volume is higher (Humblytics).

Why test on revenue or profit per visitor instead of conversion rate?

Because the tests that matter most for supplement brands change order value and margin, not only conversion. A deeper subscription discount can raise conversion rate and lose money. A three-bottle default can lower conversion rate and make more.

1. Revenue per visitor (conversion rate x average order value) for layout and page tests.
2. Gross profit per visitor for any test that changes price, discount, gifts or shipping cost.
3. Subscription take rate and early retention for subscription offer tests, checked again 60 to 90 days later.

For how this plays out in buy box and discount tests, see what subscribe and save discount to offer.

How we run Shopify A/B tests at Succession

We use Intelligems for page and offer tests, and we set up every test the same way so the result means something. Most of the work happens before the test launches.

1. Clean the baseline first. We filter to one market, usually the US, and read the funnel per landing page and per device from Shopify's funnel reports and ShopifyQL: cart rate, cart-to-checkout and checkout completion, each divided by the step before it. That tells us which step the test should target.
2. Test the big lever. If checkout completion is the weak step, we test the offer, not the layout. In our audits the biggest leaks are offer-level, such as subscription-only buy boxes and skipping the cart. When two different templates with the same offer already convert at almost the same rate, a third layout test is wasted traffic.
3. Split full pages, not fragments. Page tests run as split redirects between full page variants, so each arm is a complete experience a shopper could land on from an ad.
4. Exclude crawler sessions on new variants. A new page built for Meta collects review-crawler sessions after ads are created or edited: near 100% bounce, zero add to carts. We make sure those windows are excluded, or the new arm starts behind for reasons that have nothing to do with shoppers.
5. Read conversion rate first, then the money. Conversion rate is our first read. Before we ship a winner we confirm it on revenue per visitor, or gross profit per visitor when the test changes price, discount or gifts, as in the section above. We do not decide on ROAS, because ad spend sits in the denominator of every ROAS ratio and moves for reasons the page does not control.
6. Wait for the sample. We do not call a test before the sample supports it. When we compare periods outside a test, we use volume-matched days, because conversion rate moves when spend and traffic mix move.

For supplement brands, the tests that most often deserve the traffic are offer tests: subscription discount, a visible one-time option, the bundle default. Before testing anything, find the leaking step with add to carts but no purchases, and set expectations with a good conversion rate for a supplement store.

How long should you run a Shopify A/B test?

Until you reach the sample size you planned, and never less than one full week. Two weeks is a sensible minimum for most stores so both arms see paydays, weekends and ad changes.

• Fix the sample size before launch using the table above, and stop only when you reach it.
• Do not stop on the first significant day. Checking results daily and stopping at the first green light produces false winners.
• Avoid launching during sales events unless the sale is what you are testing.
• Keep ads stable. A big creative change mid-test shifts the audience on both arms and muddies the read.
• Segment after, not during. Look at mobile vs desktop and new vs returning once the test ends, as a source of new hypotheses, not to rescue a loser.

If you want help setting up a testing program sized to your traffic, see our services or start with a free growth audit.

FAQ