Most A/B testing advice tells you that you should test, then leaves you staring at a blank product page wondering what to actually change. This is the missing list: fifteen product page tests, ranked by the ratio of effort to likely impact, each with a hypothesis you can steal and a metric to measure. Start at the top and work down.

How this list is ranked

Two axes. Effort covers how long the variant takes to build in your theme and how much new material (photography, copy, video) it needs. Impact is an honest estimate of how often this type of test produces a meaningful result, based on practitioner experience, not invented percentages. Nobody can tell you a specific test will lift add-to-cart rate by X%; anyone who does is selling something.

For every test, the primary metric is add-to-cart rate unless noted, it’s the highest-volume signal a product page change can plausibly move, and the reasoning is covered in our complete guide to Shopify A/B testing. Before running anything, check how long your traffic level needs with a sample size calculator; low-traffic stores should prioritise the bigger, bolder changes on this list because subtle tweaks take too long to read.

One practical note: on Shopify, the clean way to run most of these is as an alternate product template (a template suffix like product.variant-b) rather than hacking JavaScript into your live page, the mechanics are in our walkthrough on how to A/B test Shopify product pages.

Low effort, high impact, start here

1. Shipping and returns info near the buy button

Effort: low · Impact: high. Take the delivery cost, delivery estimate, and returns policy that currently live in your footer or FAQ, and surface a one-line version directly under the add-to-cart button (“Free UK delivery · 60-day returns”).

Hypothesis: visitors hesitate to add because delivery cost and return risk are unknown; answering both at the moment of decision increases adds.

Measure: add-to-cart rate. If your platform shows it, also watch whether checkout completion holds up, you’re pulling forward information people previously discovered at checkout, so the effect can appear there too.

2. Social proof position

Effort: low · Impact: high. Most stores have reviews; far fewer surface the star rating and review count next to the product title and price, where the buying decision actually happens. The variant: move (or add) the rating summary above the fold, linking down to full reviews.

Hypothesis: proof at the point of decision reduces perceived risk more than proof buried three screens down.

Measure: add-to-cart rate. This pairs well with the broader set of credibility tests in our guide to trust signals on product pages.

3. Above-the-fold order on mobile

Effort: low–medium · Impact: high. On many themes, a mobile visitor sees: image, title, six trust badges, a marketing banner, then price and button. The variant: ruthlessly reorder so image, title, price, rating, and add-to-cart button all fit in the first screen or immediately after.

Hypothesis: every scroll between arrival and the buy button loses a slice of mobile visitors; compressing the decision zone increases adds.

Measure: add-to-cart rate, segmented by device, this test is really a mobile test, and desktop results will dilute it. More on this in mobile product page optimization.

4. CTA copy

Effort: trivial · Impact: moderate. “Add to cart” vs “Add to bag” vs “Get yours” vs category-specific copy (“Start my plan”). Cheap to test, occasionally surprising, rarely transformative on its own.

Hypothesis: copy that matches the customer’s mental model of the purchase lowers friction fractionally.

Measure: add-to-cart rate. Because the expected effect is small, this test needs more traffic than most, be patient or skip it if you’re low-volume. We cover button experiments in depth in CTA button A/B tests.

5. Sticky add-to-cart bar on mobile

Effort: low–medium · Impact: moderate–high. A slim bar with price and an add button that stays visible once the visitor scrolls past the main CTA.

Hypothesis: visitors who scroll deep into reviews and descriptions convert better when the buy action travels with them instead of requiring a scroll back up.

Measure: add-to-cart rate on mobile. Watch that the bar doesn’t cover other tap targets, a badly placed sticky bar can lose more than it gains.

Medium effort, high potential

6. Description format: wall of text vs scannable structure

Effort: medium · Impact: high for considered products. Rewrite one description from paragraphs into a benefit-led structure: short opening line, three to five benefit bullets, then specs in a table.

Hypothesis: visitors scan rather than read; a scannable structure gets the persuasive content actually consumed, increasing adds.

Measure: add-to-cart rate. This is a big enough topic that we’ve written a full guide to product description A/B tests.

7. Lifestyle vs studio lead image

Effort: medium (needs photography you may already have) · Impact: high. Swap the first gallery image: clean white-background product shot vs the product in use, in context, on a person or in a room.

Hypothesis: the lead image sets the emotional frame; context imagery helps visitors imagine ownership and increases adds, or, for spec-driven products, clarity beats context. Genuinely uncertain either way, which is exactly what makes it a good test.

Measure: add-to-cart rate. Full treatment, including image count, ordering, and video, in our guide to product image A/B tests.

8. Size guide prominence (apparel and footwear)

Effort: medium · Impact: high in-category. Variant: an inline size guide or “Find your size” link directly beside the size selector, instead of a link buried in the description or footer.

Hypothesis: fit uncertainty is the top hesitation in apparel; answering it at the selector increases adds (and should reduce returns over time, though a page test won’t measure that directly).

Measure: add-to-cart rate; if you can, watch size-guide engagement as a secondary signal.

9. Bundles and “complete the set” offers

Effort: medium · Impact: moderate–high on average order value. Variant: add a bundle option (“Add both and save 10%”) or a frequently-bought-together block near the CTA.

Hypothesis: a relevant bundle raises revenue per visitor without hurting the base add-to-cart rate.

Measure: this is the rare test where add-to-cart rate alone isn’t enough, you’re aiming at order value. Watch ATC rate to confirm no harm, and evaluate the bundle’s take-up separately.

10. Urgency and scarcity, done honestly

Effort: low–medium · Impact: moderate, with a warning label. Honest urgency means surfacing true information: real stock levels when genuinely low (“3 left”), real dispatch cutoffs (“Order by 2pm for same-day dispatch”), real end dates on promotions. It does not mean fake countdown timers that reset on refresh, those can produce short-term lifts while training customers to distrust everything else on your page, and they age terribly.

Hypothesis: true deadline or scarcity information helps genuinely interested visitors decide now rather than drift away.

Measure: add-to-cart rate. If a dishonest-feeling variant wins, think hard before shipping it, a page test can’t measure brand damage.

Higher effort, worth it when the basics are done

Effort: high (production) · Impact: high for demonstration-friendly products. A 20–40 second video as the second gallery item: the product in use, handled, worn, or assembled.

Hypothesis: motion answers questions statics can’t (scale, texture, mechanism), increasing purchase confidence and adds.

Measure: add-to-cart rate, and page speed. Video is heavy, and a variant that wins on persuasion can lose on load time; see A/B testing, page speed and flicker for why speed differences can contaminate test results.

Effort: medium–high (theme work) · Impact: moderate. Classic thumbnail-strip gallery vs a vertically stacked image feed (common on fashion stores) vs swipe-first on mobile.

Hypothesis: the layout that gets more images actually seen produces more confident buyers.

Measure: add-to-cart rate; image engagement as a secondary signal if available.

13. FAQ block on the product page

Effort: medium (the work is writing honest answers) · Impact: moderate–high for products that generate pre-sale questions. Variant: a five-to-eight question FAQ accordion below the description, answering the questions your support inbox actually receives.

Hypothesis: unanswered questions are silent exits; answering them on-page converts the hesitant without a support round-trip.

Measure: add-to-cart rate. Bonus: the answers double as long-tail SEO content.

14. Long-form persuasion page vs standard template

Effort: high · Impact: high variance. For a hero product, build a genuinely different template: story-driven sections, founder note, comparison table, press mentions, generous imagery, landing-page style, and test it against your standard product template.

Hypothesis: for considered or novel products, a page that fully makes the case outperforms a template designed for catalogue browsing.

Measure: add-to-cart rate. This is the kind of big-swing test that template-level testing was made for, and it’s the right move for lower-traffic stores that need large effects to reach significance.

15. Variant picker UX

Effort: medium–high · Impact: moderate, occasionally large when the current picker is bad. Dropdowns vs swatch buttons; showing which combinations are out of stock vs letting people discover it after selecting.

Hypothesis: friction and dead-ends in variant selection lose ready-to-buy visitors; a clearer picker recovers them.

Measure: add-to-cart rate, segmented by device, picker problems are usually worst on mobile.

Running these properly

Three rules that separate a testing programme from redecoration:

  1. One change per test. If the variant has new images and new copy and a moved button, a win teaches you nothing reusable.
  2. Decide the metric and the runtime before you start. Peeking at results daily and stopping when you like the number is how false positives get shipped.
  3. Consistent visitor assignment. If the same visitor sees version A on Monday and version B on Wednesday, your data is contaminated and your customers are confused.

That third point is where tooling matters. Atchoo! handles it by assigning each visitor deterministically to a variant via a first-party cookie, the same person always sees the same version, while serving variants through Shopify’s native template mechanism, so a “variant” is just another product template in your existing theme. It tracks add-to-cart rate per variant in real time with a traffic-source breakdown, and calls winners with Bayesian analysis, which gives you “there’s a 96% probability B is better” instead of a p-value you have to interpret.

Frequently asked questions

Which product page test should I run first?

The cheapest test with a plausible large effect for your store. For most merchants that’s shipping/returns visibility near the buy button (#1) or social proof position (#2), both are quick theme edits with a credible mechanism behind them. If your mobile page buries the CTA, the above-the-fold reorder (#3) jumps the queue.

How long should a product page A/B test run?

Until it reaches the sample size you calculated up front, typically two to six weeks for stores with moderate traffic, and ideally over whole weeks, so weekday/weekend behaviour is balanced. Fixed calendar rules (“always run 30 days”) are less useful than a proper sample size and duration calculation.

Can I test more than one idea at the same time?

On different products or templates, yes. On the same page, avoid it, overlapping changes make results unreadable. The exception is a deliberate “big swing” test (#14) where the variant is intentionally a bundle of changes and you accept that you’re testing the concept, not the components.

What if a test shows no difference?

That’s a real result: it means the element you changed doesn’t matter much to your visitors, and you can stop debating it internally. Log it, keep whichever version you prefer for other reasons, and move to the next idea. A testing programme’s value comes from the accumulated map of what matters, not from every test winning.

Do these tests work for low-traffic stores?

The high-impact, big-change ones do (#1, #3, #7, #14). Subtle tests like CTA copy need volumes small stores don’t have. Our guide to A/B testing for low-traffic stores covers how to choose accordingly.

Ready to run one? Atchoo! turns any alternate product template in your theme into a live A/B test, 50/50 traffic split, real-time add-to-cart tracking, Bayesian winner calls. Full testing is on the Pro plan at $39/mo with a 14-day free trial. See pricing or start with the complete guide to Shopify A/B testing.