The 2026 Ecommerce CRO Checklist: 40+ Tests That Actually Move Revenue

Across 47 D2C audits we ran in the last 12 months, the median Shopify store was leaving 23 to 31 percent of its potential revenue on the floor. Not on ads. Not on catalog. On the funnel itself. Same traffic, same products, same margins, better structure.

This is the checklist we run every new engagement through. It is ordered by where money usually hides, not by where CRO blogs tell you to look.

→ Most stores fix cart and checkout first. Start with the product page. → A 40-item checklist is useless without priority tiers. Use IMPACT 5 below. → Benchmarks included are from real Shopify and WooCommerce accounts in the $500k to $18M ARR range. → The goal of this post is a concrete list of tests, not another "improve trust" sermon.

Why this checklist exists

Most CRO content is written for enterprise retailers with analysts, session replay tools, and budget for 6-month test roadmaps. Real D2C operators have a Shopify theme, a Klaviyo account, a weekly ad spend they have to justify, and a founder who wants more orders by Friday.

So we built this around three rules.

First, every test has to have a plausible lift band we have seen in production. No "this usually works" hand-waving. If we cannot point to at least 5 instances of it hitting, it is not on the list.

Second, every test has to be scoped so a single person on a small team can ship it inside one sprint. No "rebuild the checkout from scratch" vanity tests.

Third, the checklist is organized by page, not by theme. "Increase trust" spans the whole site. "Fix the PDP sticky ATC on iOS" is a ticket.

If you want the deeper reasoning behind product pages specifically, read our breakdown on product page CRO patterns. If you want the counterintuitive tests that beat "best practice" defaults, we wrote those up in ecommerce CRO tests that beat best practices. This post is the superset, the thing you hand a dev on day one.

The IMPACT 5 priority framework

We label every test with one of 6 priority tiers. We call it IMPACT 5 because 5 of the 6 tiers describe shippable work and the 6th is a catch-all for edge cases.

→ I — Immediate. Ship this week. Effort is under 1 day. Lift is proven across 20+ stores. No design review needed. → M — Medium. Ship this sprint. Effort is 2 to 5 days. Requires design or content work. Lift is proven but context-dependent. → P — Projecting. Ship next quarter. Effort is 1 to 3 weeks. Lift is larger but slower to realize (repeat purchase, LTV). → A — Aligning. Cross-functional. Needs marketing, ops, or CX alignment. Often involves policy changes (returns, shipping, subscription). → C — Consistent. Ongoing hygiene, not a one-time ship. Think inventory feed cleanliness, review moderation, bounce investigations. → T — Tailored. Category-specific. Works for apparel but not supplements, or vice versa. Flag for context before testing.

Priority is not difficulty. Priority is the expected path from ticket to revenue. Easy tests can be low priority if the lift is small. Hard tests can be high priority if the money is sitting there.

Quick reference table

These are not guarantees. They are the ranges we see. If your store is already optimized on one dimension, the lift from that test will collapse toward zero.

PDP wins (high impact)

The product detail page is where most ecommerce revenue is won or lost. A 1 percent PDP conversion lift on a store doing $3M in ARR is worth roughly $30k incrementally. A 1 percent checkout lift on the same store is worth maybe $8k because far fewer sessions reach checkout. Start here.

1. Sticky add-to-cart bar on mobile (I) Persistent bar at the bottom of the viewport showing product name, variant, price, ATC button. Still one of the single highest-ROI changes on any Shopify store that does not have it. Expect 6 to 14 percent mobile conversion lift. Dawn-based themes ship this now; Debut-era themes do not.

2. Lifestyle image as hero, not product-on-white (I) Product-on-white belongs in the gallery, not the hero slot. Lifestyle shots (product in use, product in hand, product in context) lift conversion 5 to 10 percent on almost every audit we run. Keep the product-on-white, just move it to position 2.

3. Show reviews above the fold (I) Star rating and count, visible without scrolling. Clickable to reviews section. If you have fewer than 20 reviews per SKU, aggregate category-level social proof instead ("4.8 stars across 2,100 reviews").

4. Price and compare-at together (I) "$89 $129" with strikethrough on the higher number. Do not fake compare-at prices. Shopify stores that use compare-at prices without real reference prices are getting flagged in multiple jurisdictions now.

5. Free shipping threshold message near price (I) "Free shipping over $75" or "You are $18 from free shipping." Surprise shipping cost is still the #1 checkout abandonment trigger in every checkout survey published since 2018.

6. Variant pills, not dropdowns (I) For up to 8 options, show swatches or pills. Dropdowns hide information. Pre-select the most popular variant so the price is always visible.

7. Out-of-stock with notify-me (M) Grey out the variant, keep it visible, swap the ATC button for "Notify me." Collects email, saves the session. Every major ESP integrates this now.

8. Review photos surfaced in gallery (M) UGC alongside product photography. Not in a separate "reviews" section 8 scrolls down. Lifts PDP conversion 5 to 13 percent on apparel, home goods, and beauty.

9. Benefit bullets, not feature bullets (I) "Dishwasher safe" beats "Ceramic construction." Three max, above the fold. Long feature lists belong in the description.

10. Size guide modal, not separate page (I) Tap to open, size chart overlays the PDP. Do not send users to a dedicated /size-guide page; they rarely come back.

11. Shipping and return policy in a tab or accordion (M) On the PDP, not 3 clicks away. Accordions below the ATC work. Dedicated tabs below the gallery work. Links to a separate policy page do not.

12. Subscription toggle if you have repeatable consumables (P) Supplements, coffee, beauty, pet food. One-time vs subscribe-and-save toggle directly under the price. Lifts 90-day LTV 8 to 20 percent when executed well.

13. Video only if the product benefits from it (T) Jewelry sparkle, candle burn, food prep, mechanical motion. Auto-playing video on a static product (bedsheets, mugs, notebooks) is mostly friction.

Collection and filtering

Collection pages are the forgotten middle of the funnel. They are where paid traffic often lands and where organic browsing happens. A well-built collection page compounds every PDP improvement downstream.

14. Best-selling default sort (I) "Featured" sort is almost always theme default and almost always wrong. Sort by best sellers or manual curation. Never by price ascending, which buries margin.

15. Filters on the left or collapsible top on mobile (M) Facets matter more than they used to. Buyers expect to filter by size, color, price, availability, and category-specific attributes. If your filter UX is hidden or broken on mobile, collection conversion collapses 15 to 30 percent.

16. Color swatches on collection tiles (M) Hover or tap on a tile swaps to the selected color. Saves a round trip to the PDP for users who just want to see options. Particularly high ROI for apparel and home.

17. Quick add to cart on collection tiles (T) For single-variant or minimal-variant products (candles, books, food). Skip the PDP entirely. For multi-variant products (apparel with size), quick add hurts because users forget to pick the variant.

18. Show in-stock count for low-stock items (I) "Only 3 left in size M." Do not fake it. If you fake scarcity and users catch you (they do), you lose the sale and the review.

19. Collection page heading with category context (M) Not "Tops." Something like "Tops made from 100 percent organic cotton, ethically sourced." Helps SEO, orients the buyer, sets expectations. This is the same pattern we recommend in the D2C ecommerce SEO guide.

20. Pagination or infinite scroll, not both (I) Pick one. Both is confusing. For catalogs under 60 items, show all on one page. For 60 to 200, paginate. For 200+, add search.

21. Breadcrumbs on collection and PDP (I) Not for aesthetics. For orientation and navigation back up the tree. Also JSON-LD breadcrumb schema for SERP display.

22. Collection banner images that do not push product below the fold (M) Keep collection heroes under 200px on mobile, 300px on desktop. Every pixel of hero image is a pixel of product the user is not seeing.

Cart and checkout

Checkout optimization is where the industry tests most and learns least. Most "best practices" do not apply to your store. The following are the ones that consistently do.

23. Express checkout above the fold on cart (I) Apple Pay, Shop Pay, Google Pay, PayPal. Visible before the user has to scroll. 8 to 17 percent lift on mobile cart conversion when added to stores that only had them at checkout.

24. Guest checkout as default, account as opt-in (M) Forcing account creation before checkout reliably kills 10 to 22 percent of first-time buyer conversion. Collect the email, let them buy, prompt for account creation post-purchase.

25. One-column checkout, not two (M) Two-column checkout with cart summary on the right looked modern in 2019. One-column with a collapsible order summary on mobile now outperforms on almost every audit.

26. Autofill address via API (M) Google Places or Loqate or Shopify Shop Pay autofill. Typing a full address is a 40-second task on mobile that users fail at a 12 to 18 percent rate. Autofill it.

27. Show total cost before payment step (I) Shipping, tax, discount all visible before the user enters a card. Surprise cost at the final step is the single most tested and most validated abandonment trigger.

28. Progress indicator across checkout steps (I) "Information → Shipping → Payment." 3 steps, visually clear, not 7. Progress indicators reduce friction perception even when they do not reduce actual friction.

29. Error messaging inline, not at submit (I) Invalid zip, invalid card, missing field. Tell the user at the moment of error, not when they hit "Place Order." This is a 30-minute dev change that often lifts checkout completion 2 to 5 percent.

30. Discount code field collapsed, not open (M) Open discount field prompts users to hunt for codes. They leave, search, come back, or do not come back. Collapse it behind "Have a discount code?"

31. Smart cart upsells, not random ones (P) "Pairs well with" based on actual co-purchase data. Not "You might also like" based on category alone. When done well, 3 to 9 percent AOV lift.

32. Sticky cart drawer, not cart page redirect (M) Adding to cart should open a slide-over drawer, not navigate to /cart. The redirect breaks browsing momentum. Drawer keeps it.

33. Exit-intent on cart page only (T) Not site-wide. Cart-page exit-intent with a targeted offer (free shipping, 10 percent off) recovers 2 to 5 percent of would-be abandoners without cannibalizing full-price buyers.

34. Address validation before submit (M) Catches typos and undeliverable addresses. Reduces "address correction" customer service tickets by 30 to 60 percent and saves you shipping-carrier surcharges.

35. Failed payment recovery (A) Involuntary churn on subscriptions and failed first-purchase payments. Dunning, smart retries, card-updater services. This is revenue you already earned; go get it.

Post-purchase loops

Post-purchase is where most stores stop optimizing. It is also where LTV economics hide. The D2C brands with the best unit economics are often the ones that spent the last 6 months fixing order confirmation, review solicitation, and second-purchase timing.

36. One-click upsell on order confirmation (M) "Add to your order" on the thank-you page, before payment settles. Apps like ReConvert, AfterSell. 3 to 9 percent AOV uplift when curated well, near zero when lazy.

37. Branded order tracking page, not carrier's page (M) Shipment tracking on your own domain, with upsells and content. Users check tracking 3 to 8 times per order on average. That is attention you are currently giving to UPS.

38. Review request timed to product experience (A) Not 3 days after order. When the product has been used. For apparel, 10 to 14 days. For consumables, 21 to 30 days. For furniture, 30 to 60 days. Product-experience timing lifts review response rate 2 to 4x.

39. Second-purchase email at category-specific interval (A) If 40 percent of your buyers repeat within 45 days and you are sending the "come back" email at day 14, you are wasting sends. Pull your actual repeat curve and build around it.

40. Referral offer in post-purchase flow (P) Not on the thank-you page directly (premature). In the shipped-email or delivered-email moment, when the customer has just received the product. Referral conversion doubles in the delight window.

41. Subscription modification, not just cancellation (A) Pause, skip, swap, downgrade. If the only button is "cancel," users cancel. If pause and swap are equally prominent, churn drops 15 to 30 percent on recurring products.

42. Win-back sequence at lapse, not at 180 days (A) If a customer normally reorders at day 30 and you start win-back at day 90, you are late. Define lapse as 1.5x your repeat-purchase interval.

43. Refund and return self-service (A) Self-service returns portal. Every support ticket avoided is both margin and customer satisfaction. Required for any store over $2M ARR.

Measurement

Tests without measurement are theater. The following are the measurement disciplines that separate teams that compound from teams that guess.

44. Track add-to-cart, checkout, and purchase as 3 distinct events (C) Shopify theme tracking often conflates these. GA4 and your attribution tool need separate events. If you cannot see cart-to-checkout drop-off, you cannot fix it.

45. Minimum 14-day test windows (C) Shorter windows are noise. Weekend traffic differs from weekday. Payday weeks differ from mid-month. 14 days catches one full cycle minimum. Longer for lower-volume stores.

46. Pre-register hypothesis, expected lift, and primary metric (C) Before the test ships. In writing. In a shared doc. Teams that do not do this end up inventing post-hoc explanations for every result.

47. Decide on a primary metric, live with it (C) If your primary metric is order conversion rate and the test wins on AOV but loses on conversion, you did not win the test. Pick one metric, stick with it, then evaluate secondary metrics.

48. Segment results by device, by traffic source, by new-vs-returning (C) Average lift hides real patterns. A test that is +8 percent on mobile but -3 percent on desktop is a mobile win, not a site-wide roll-out.

What to do this week

→ Install the mobile sticky ATC if you do not have one. This is often the single largest unconstrained lever on a Shopify store. → Move express checkout buttons (Apple Pay, Shop Pay, PayPal) above the fold on your cart page. → Switch guest checkout on as the default. Make account creation post-purchase, not pre-purchase. → Pull your actual repeat-purchase curve from the last 12 months and compare it to your email flow timing. Realign whichever is off. → Audit one collection page for filter UX on mobile. If it is hidden or broken, that is your first test.

Common traps

Testing a theme swap as "CRO." Swapping your whole theme is not a test; it is 30 tests at once with no attribution. Ship themes in increments that can be attributed.

Trusting Shopify's built-in "conversion rate" number. It includes bot traffic, conflates device types, and rolls up orders regardless of source. Build your own dashboard against GA4 or your attribution tool.

Optimizing the checkout before the PDP. Most stores push on Shop Pay and one-click upsells before they have fixed basic PDP hygiene. Fix the top of the funnel first; every downstream win compounds.

Shipping "trust signals" as a category. Badges, guarantees, press logos. These only lift conversion on stores that genuinely lacked trust signals. On stores that already had them, adding more is noise.

Testing copy changes without clear hypotheses. "Let's test a new headline" is not a test. "Let's test a benefit-led headline against a feature-led headline because our audience skews to first-time buyers" is a test.

If you want someone to run this program for you, that is what our CRO practice does. If you want it paired with organic growth, we bundle with SEO.

FAQ

How long should I run each of these tests? Minimum 14 days and minimum 300 conversions on the primary metric. Shorter than that and you are reading noise. On lower-volume stores ($500k ARR and below), plan for 3 to 4 week test windows.

What percentage lift should I expect on a fully optimized store? On stores that have already done most of this work, further tests usually yield 1 to 3 percent lifts, not 10 to 20 percent. That is still worth doing: a 2 percent lift on an $8M store is $160k a year. But calibrate expectations by where you are in the curve, not where a case study is.

Should I use a CRO tool like VWO or Optimizely on a small store? Below $2M ARR, Shopify's built-in A/B capabilities and Google Optimize alternatives are usually enough. Above $2M, a dedicated CRO platform starts paying for itself in test velocity and segmentation.

Does any of this change for subscription vs one-time purchase? Subscription stores should weight items 12, 35, 41, and 42 higher. Subscription churn and payment recovery are where most of the money hides. One-time purchase stores should weight post-purchase upsell (36) and referral (40) higher because there is no recurring revenue to compound.

Our product category is niche. Will these still work? About 80 percent will. The other 20 percent (marked with T for Tailored in the framework) are category-sensitive. The best way to know is to test in priority order and trust what your traffic tells you over what a benchmark claims.