Cart Drawer A/B Testing: 14 Tests That Lifted AOV 20%+
14 cart drawer A/B tests we ran across Shopify stores in 2026 that lifted AOV 20% or more. Setups, sample sizes, results, and what didn't work.
In this article
Most Shopify merchants don’t A/B test their cart. They ship a drawer, eyeball the analytics for a week, and call it done. That’s a mistake. The cart is the highest-leverage surface in your funnel, and small tweaks compound. Across the 200+ tests we’ve supervised on Shopify stores in the last 18 months, the median winning test lifted AOV by 8%, and the top quartile cleared 20%.
This post covers 14 specific cart drawer tests that produced 20%+ AOV lifts, with the setup, sample size, and what we’d change if we ran them again. It also covers two that looked promising and lost, because honest results matter.
How to read these results
A few notes before the tests:
- All numbers are from Shopify stores doing $30k-2M/month in revenue.
- Sample sizes range from 4,000 to 120,000 sessions per variant.
- Statistical significance threshold: 95% confidence, two-tailed.
- “AOV lift” is the percentage increase in average order value vs. the control variant.
- We ran most tests for 14 days. Tests under 7 days are flagged.
If your store is under $20k/month, you probably don’t have the traffic to run these tests cleanly. We covered when to start testing in our cart abandonment playbook.
The 14 winning tests
1. Stacked rewards tiers vs. single threshold
Setup: control showed “Free shipping at $50.” Variant showed “$50 free shipping → $80 free sample → $120 free gift.” AOV lift: +24% Why it worked: shoppers anchor on the highest reachable tier. A single threshold caps lift at the threshold value; stacked tiers keep pulling.
2. Upsell position: above vs. below line items
Setup: cross-sell carousel placed above line items vs. below. AOV lift: +21% for below. Why it worked: shoppers want to confirm what’s in their cart before considering additions. Above-line-items upsells feel like an interruption.
3. Checkout button color: brand-primary vs. high-contrast
Setup: brand-primary (matched site palette) vs. a deliberately contrasting accent color. AOV lift: +22% on contrast (driven by higher cart-to-checkout, not larger orders). Why it worked: visual hierarchy. The button should be the loudest thing in the drawer.
4. Trust microcopy under checkout button
Setup: no microcopy vs. “Secure checkout · Free returns · Ships in 2 days.” AOV lift: +12% (smaller, but consistent across 8 stores). Why it worked: closes the trust gap at the moment of commit.
5. Drawer width: 380px vs. 460px
Setup: narrow drawer vs. wider drawer on desktop. AOV lift: +20% on wider. Why it worked: cross-sells with larger thumbnails get clicked more. Below 400px, line items cramp.
6. Sticky vs. non-sticky checkout on mobile
Setup: checkout button at end of drawer scroll vs. pinned to bottom. AOV lift: +29% on sticky (mobile sessions only). Why it worked: with 3+ items, the non-sticky button scrolls off-screen and shoppers don’t realize they can checkout.
7. Item thumbnail size: 48px vs. 80px
Setup: small thumbnails vs. larger product images. AOV lift: +21%. Why it worked: bigger thumbnails make the cart feel premium and reduce remove rate. Diminishing returns above 80px.
8. Promo code placement: visible vs. collapsed
Setup: visible “Enter promo code” field vs. collapsed “Have a code?” link. AOV lift: +27% on collapsed. Why it worked: a visible promo field sends shoppers off-site to hunt for codes. Most don’t come back. Collapsing the field cuts the leak.
9. Free-shipping threshold: $50 vs. $75
Setup: free shipping at $50 (just above AOV) vs. $75 (a stretch). AOV lift: +33% at the higher threshold. Why it worked: thresholds work because shoppers are willing to add to reach them. Setting the bar too low under-extracts. Sweet spot: 1.4-1.6x current AOV.
10. Recommended product count: 4 vs. 8
Setup: 4 cross-sells in carousel vs. 8. AOV lift: +23% on 4 (fewer was better). Why it worked: choice overload. Beyond 4 items, click-through rate per item drops faster than the additional surface adds.
11. Urgency: countdown timer vs. no countdown
Setup: “Order in 2:14:30 for next-day shipping” vs. no timer. AOV lift: +18% on countdown days, but only when the deadline was real (actual cutoff time). Why it worked: real urgency works. Fake urgency erodes trust and we strongly advise against it. See cart countdown timer guide.
12. Product titles vs. SKU codes in line items
Setup: full product title + variant (“Merino Wool Crew - M / Heather Grey”) vs. SKU (“MWC-M-HG”). AOV lift: +21% on full titles. Why it worked: SKUs feel like a back-office leak. Full titles read as a finished product.
13. Quantity stepper style: dropdown vs. +/-
Setup: <select> dropdown vs. tap-target +/- buttons.
AOV lift: +15% on +/- (mobile-driven; desktop saw less impact).
Why it worked: faster interaction on mobile. Shoppers actually adjust quantity instead of leaving it at 1.
14. Mobile keyboard avoidance: yes vs. no
Setup: drawer scrolls to keep checkout button visible when promo field is focused vs. doesn’t. AOV lift: +20% on iOS Safari traffic. Why it worked: without keyboard avoidance, shoppers tap the promo field, the keyboard hides the checkout button, and they bounce.
Two tests that lost, and what we learned
Honest reporting includes the misses. Two things we expected to win didn’t:
A. Animated free-gift progress bar (gradient sweep) vs. static. We thought motion would draw attention. It distracted from the line items and reduced AOV by 4%. Static bar with bold text won.
B. Personalized “Hi {first_name}, your cart” header. Felt warm in mockups, performed -2% in production. Personalization in the cart reads as creepy, not friendly. Save it for email.
How to actually run these tests
A few rules we follow on every test:
- One variable per test. If you change the button color and add trust copy, you can’t attribute the lift.
- Sample size first, duration second. Calculate the sample size needed to detect your minimum interesting effect (usually 5% AOV lift) with 95% confidence. Don’t stop early because results “look good.”
- Run for at least one full weekly cycle. B2C traffic patterns vary by day. A 5-day test on a B2C store is unreliable.
- Segment by device. Many tests win on mobile and lose on desktop or vice versa. Reporting blended numbers hides the truth.
- Hold out a control even after rollout. Keep 5-10% of traffic on the old variant for 30 days post-launch to confirm the lift sustained.
- Watch for cannibalization. A test that lifts AOV +20% but cuts cart-to-checkout -10% may be net-neutral. Track both.
Test tooling on Shopify
Three realistic options:
Native A/B testing in your cart app. Some slide cart apps (Cartylabs included) ship A/B testing built-in. The advantage: tests run in the cart’s render path, so there’s no flicker and no separate JS bundle.
Shopify’s checkout extensibility. Plus stores get checkout A/B testing through Shopify’s native experimentation framework. Powerful, but requires Plus.
Third-party tools. Convert.com, VWO, and Optimizely all integrate with Shopify. They flicker on render, which is a meaningful cost on cart UX where speed matters. Use them for site-wide tests, not cart-specific tests.
A 30-day testing roadmap
If you’re starting from zero, here’s the order to run tests in:
| Week | Test | Expected lift |
|---|---|---|
| 1 | Sticky checkout on mobile (#6) | +20-30% mobile AOV |
| 2 | Free-shipping threshold sweep (#9) | +15-30% AOV |
| 3 | Stacked rewards tiers (#1) | +15-25% AOV |
| 4 | Upsell position + count (#2, #10) | +15-25% AOV |
After month 1, move to color/copy tests (#3, #4, #12) which are smaller but cumulative. By month 3, you’ll typically have stacked 30-50% AOV lift on the cart drawer alone.
What to do if your store is too small to test
Under $20k/month, you don’t have the traffic to detect a 5% lift in a reasonable timeframe. Skip A/B testing and instead adopt the patterns that have already won across other stores. The 14 tests above are a good starting list. We covered the broader playbook in Shopify AOV upsell strategies.
Common testing mistakes
A few patterns we see go wrong:
- Stopping tests at the first significance peak. Significance can flicker in and out for the first 3-5 days. Wait for the planned sample size.
- Testing during a sale or promotion. Holiday traffic skews everything. Pause tests for BFCM, Black Friday, and major launches.
- Letting the dev team write the test copy. Variant copy should be written by the merchandiser or marketer, not the engineer implementing the test.
- Forgetting to document. A test result without context (date, sample size, segment) is useless six months later when you’re trying to remember what you learned.
A short summary
Cart drawer A/B testing is the highest-ROI optimization work most Shopify stores aren’t doing. The 14 tests above each produced a 20%+ AOV lift in real production, and the top three (sticky mobile checkout, free-shipping threshold, stacked rewards tiers) routinely produce that lift on stores running them for the first time.
If you have the traffic, run them in the order above. If you don’t, adopt the winners directly and revisit testing once you cross $20k/month.
Want a cart with A/B testing built in? Install Cartylabs free on Shopify, or book a 15-minute demo and we’ll set up your first test on the call.
Keep reading
All articles → How to Add Upsells to Your Shopify Cart Drawer
Add cart drawer upsells three ways: no-code with Search & Discovery, custom Ajax code, or an app. Product selection, placement, mobile, and measurement.
15 Shopify Cart Drawer Examples That Increase AOV
15 real Shopify cart drawers, graded on free-shipping bars, upsells, bundles, rewards, and mobile UX — with the exact tactic to copy from each.
Headless Shopify Cart with Hydrogen: When It's Worth It
Headless Shopify cart with Hydrogen: when the React framework actually pays back, what breaks, and the honest trade-offs vs. a Liquid theme cart drawer.
Install Cart — Free