Shopify Cart Page Optimization: A Test Log, Starting With the One That Lost
The test changed the cart on a subscription food brand's store: a rebuilt cart drawer, shipped as a new theme version and split against the old one. It looked like a sensible set of improvements, the kind most cart guides recommend.
We'll get to what happened next in case 2. First, the case that explains a lot of what goes on in a Shopify cart, and it isn't one of our tests.
Why Most Cart Advice Adds Things
Open a typical cart drawer on a phone. A free shipping banner across the top. Two lines of product. A carousel of "you may also like." A row of trust badges. Somewhere below all of it, the checkout button, and the order total is under the banner.
Each of those pieces may raise average order value on the app's dashboard. If that's the number you're after, this log won't help much. It's for stores that care about revenue per visitor, the number that counts the people who left as well as the ones who added one more item.
So every case below follows one rule: test an element before you install it for everyone.
Case 1: The Price Check
Source: Baymard Institute research, not one of our tests.
In Baymard's testing, 64% of users looked for shipping costs on the product page before adding anything to the cart. Yet 43% of ecommerce sites showed no shipping estimate there. When it's missing, the cart becomes the only way to find the total, and a shopper who adds, looks and leaves shows up in your reports as an abandoned cart.
Some of those were never going to be orders. They were questions.
What the cart should do: show the full order total, or an honest estimate including shipping and taxes. Baymard's guideline for the cart is blunt: users given only part of the total "will be reluctant to proceed from the cart and begin filling out multiple form fields." Show a delivery date as well. A range like "3 to 5 business days" leaves shoppers to work out for themselves when the order will arrive, as Baymard puts it.
What the product page should do: answer the question before the shopper needs the cart to ask it. We covered cost transparency in our checkout optimization guide.
Case 2: The Cart That Did Too Much
Source: our client test, run in Intelligems on a subscription food brand's store, from the end of August 2026. About 7,700 visitors saw the control cart and 7,300 the variant, for 142 and 110 orders.
What changed: the variant was a rebuilt cart drawer, tested as a theme split. The element list belongs to the hypothesis write-up; the point of this case doesn't depend on it.
What happened: conversion went from 1.84% to 1.50%, a 19% drop. Revenue per visitor fell 25%. Average order value fell 8%. By the time we paused it, the probability that the control was the better cart stood at 95%.
What we did first: checked the variant before doubting the idea. When a test turns red early, the first question at Weblics is whether something in the variant is broken, because a bug looks exactly like a bad idea in the numbers. We went back through the variant cart on phones and desktop, compared it against the design, and checked every step to checkout.
Then one step further. A theme split can go wrong in ways a design review won't show, so we paused the cart test and ran an A/A test on the same mechanism: two identical themes, a 50/50 split, two weeks. Conversion came out at 1.06% against 1.07%, revenue per visitor 3% apart, on 15,066 and 14,924 visitors and 160 and 159 orders. Flat, as it should be. The split was clean. The loss was the cart.
What we learned: a cart with more in it gave people more to decide at the moment they had already decided. Each added element was a reasonable idea on its own. Together they cost a fifth of the orders.
The general point: a cart change can hurt more than the cart. Revenue per visitor fell further than conversion did, which means the people who still bought also bought less. That's why we read cart tests on revenue per visitor, across the whole visit.
Case 3: The Add-On That Didn't Hurt
Source: our client test, run in Intelligems on a pasta brand's store, 16 to 22 September 2026. About 11,000 visitors per version, 449 orders on the control and 471 on the variant.
What changed: single packs offered in the mini cart at $8.99 instead of $9.99, one dollar off, with the catalog price unchanged everywhere else.
What happened: conversion +5%, revenue per visitor +9%, average order value +4%. The probability that the variant was ahead sat at 78 to 82%, short of where we call a winner. It ran its planned six days and stopped.
How to read it: the direction is the useful part. An add-on that lowered the price of the extra item did not lower the value of the visit, which is the fear with cart discounts. A longer run would settle whether it raised it. What we'd not do is read the AOV line on its own and call it a win.
Case 4: The Free Shipping Line
Source: our client test, run in Intelligems on a Dutch apparel brand's store from late July 2026, and the same brand's post-purchase survey.
Why it matters: in Baymard's latest abandonment survey, 40% of shoppers who abandoned an order, leaving aside those just browsing, said extra costs like shipping, tax and fees were too high. It's the biggest single reason on the list.
What changed: the free shipping threshold moved from €45 to €50 in one variant and to €55 in another, everywhere the store shows it.
What happened: at €50, conversion +12%, revenue per visitor +19%, average order value +7% and gross profit per visitor +20%, on about 7,900 visitors on the control and 7,600 on the variant, 279 and 301 orders. The probability the variant was ahead sat at 91 to 94%. At €55, conversion +8% and revenue per visitor +12%. The test was paused before either reached our bar, so read both as direction.
What the numbers missed: around the time the test ran, the brand's post-purchase survey asked about 800 buyers what almost stopped them. One of the top answers was the €50 line itself, which sat a cent above the store's €49.99 bundle price. Buyers called it manipulative. Revenue per visitor said the threshold worked. The survey said some of the people it worked on resented it. A threshold test needs both readings, and a line that lands a euro above your best-selling price is a decision, not a rounding.
How to read your own: one variant, one threshold, read on revenue per visitor after the cost of the shipping you now give away. Average order value alone will usually look good, because the threshold pushes some baskets over the line. Our checkout guide covers how to pick a starting threshold.
Case 5: The Link That Left
Source: found during one of our cart reviews on a client store and removed. We didn't A/B test the removal, so there is no number for what it was costing, and we're not going to invent one.
This one is written up in full in where to add reviews on Shopify. The short version for the cart: nothing in it should open another page. In Baymard's testing, 53% of users actively sought out negative reviews anyway. A link under the checkout button helps them find those reviews at exactly the wrong moment.
Case 6: The Coupon Field
Source: Baymard Institute research, not one of our tests.
Baymard lists it among the frictions that analytics alone won't show you: a discount code field that draws shoppers away to search Google for coupons. They leave, find a code site, maybe find a code, maybe don't come back.
Shopify's checkout has its own discount code field. If your theme or a cart app adds a second, prominent one in the cart, that's a candidate for your first test: collapse it behind a small link, or remove it from the cart and leave the one in checkout.
How we'd test it: one variant, one change, read on revenue per visitor. If you run a lot of promotions, we'd expect this to matter less, since your shoppers already have codes.
When the Cart Isn't the Problem
Take one of the most quoted numbers in cart abandonment. In Baymard's latest abandonment survey, 18% of shoppers abandoned an order because the site wanted them to create an account. On Shopify, that's a single setting: "Require customers to sign in to their account before checkout", under Settings > Checkout. Leave it off and shoppers can check out as guests. Shopify's help page also notes that when it's on, "accelerated checkout options such as Apple Pay aren't displayed in the online store cart." Check it before you redesign anything.
Then look at where your cart visitors came from. If many of them add to cart straight from a product page and leave within seconds, a good share were likely checking the price. Put shipping information on the product page, and some of those visits either stop or turn into orders. Where that line sits on the page, and what else belongs next to the button, is covered in our Shopify product page optimization guide.
Shopify also recommends placing accelerated checkout buttons near the checkout button in the cart, so shoppers can choose between the regular checkout and a faster one. Make sure yours are there, and that nothing sits between them and the thumb.
That leaves three ways to handle your cart. You can leave it as it is and keep reading the abandoned cart number as a verdict. You can install the next widget and hope, the way the variant in case 2 did. Or you can change one element at a time and read each change on revenue per visitor. Only the third one tells you anything.
The people who make the third choice are the ones who can say "not until we've tested it" in a meeting and turn out to be right. It isn't a big project. One element per test, and any change can be rolled back. The shoppers in your cart are the ones you've already paid the most to bring this far.
If you'd rather not work through it alone, send us your store. We'll go through your cart the way this log does and come back with what we'd test first.
The cart isn't where you win the order. It's where you stop losing it.
Sources
• Baymard Institute, guideline #732 "Essential Price Info for the Cart" (order total guideline)
• Baymard Institute, "Use 'Delivery Date' Not 'Shipping Speed'" (41% of sites show only a shipping speed)
• Baymard Institute, cart abandonment rate statistics and reasons for abandonment during checkout (42% just browsing; 40% extra costs, 20% slow delivery, 19% trust, 18% account)
• Baymard Institute, "Respond to Negative User Reviews" (53% of users seek out negative reviews)
• Baymard Institute, "How to Audit Your Checkout Flow for Hidden Friction" (discount code field drawing users away to search for coupons)
• Shopify Help Center, checkout form options (mandatory sign-in hides accelerated checkout options such as Apple Pay from the cart)
• Shopify.dev, theme docs on accelerated checkout (place the buttons near the checkout input of the cart form)
• Weblics test log (Intelligems): subscription food brand cart test, from 31 August 2026, and the A/A test that followed it, September 2026; pasta brand mini cart price test, 16 to 22 September 2026; Dutch apparel brand free shipping threshold test, from 28 July 2026
• Client post-purchase survey (Grapevine), Dutch apparel brand, 12 July to 11 August 2026, about 805 responses
.webp)
.webp)
.webp)
.webp)
.webp)
.webp)
Do you have any questions left?
Here are the answers for you
Show the full order total or an honest estimate including shipping and taxes, show a delivery date, keep the checkout and accelerated checkout buttons easy to reach on a phone, and remove anything that links away from the cart. Then test changes one at a time and judge them on revenue per visitor, not average order value alone.
In Baymard's latest survey of US shoppers, leaving aside the 42% who were just browsing, the top reasons were extra costs (40%), slow delivery (20%), not trusting the site with card details (19%) and being asked to create an account (18%). Baymard also found 64% of users look for shipping costs on the product page, and many add to the cart just to see the total.
Yes. Baymard's research found that users who see only part of the order total are reluctant to leave the cart and start filling in checkout forms. Show the total with shipping, or a clear estimate, plus a delivery date.
Usually not a prominent one. Baymard found that discount code fields can draw shoppers away to search for coupons. Shopify's checkout already has a discount field, so a second one in the cart is worth testing collapsed or removed.
They can. Any element between the shopper and the checkout button is a decision, and in one of our cart tests a variant with added elements cut conversion by 19% and revenue per visitor by 25%. In another, a one-dollar discount on a mini cart add-on moved conversion and revenue per visitor up, though not significantly. Test upsells with revenue per visitor as the deciding number, not average order value.
Shopify recommends placing them near the cart's checkout button. Note that when "Require customers to sign in to their account before checkout" is on, accelerated checkout options such as Apple Pay aren't displayed in the cart, according to Shopify's help page.
Every plan includes complete care-driven CRO - what varies is testing capacity and analysis depth.
All Plans Include:
Onboarding (First 5 days):
- Founder interviews & business deep-dive
- Comprehensive technical website audit
- Customer psychology analysis (ICP, 5 WHYs, SWOT)
- AI-trained buyer personas creation
- Ad creatives audit
- Marketing ecosystem review
Ongoing (Continuous):
- Psychology-first hypothesis generation
- Conversion-focused UX/UI design
- Strategic copywriting
- Shopify development & implementation
- A/B testing & QA
- Transparent reporting & documentation
- Strategy meetings (weekly or bi-weekly)
What Changes by Tier:
- Tests per month: 2, 4, 6, or 8 A/B tests
- Meeting frequency: Bi-weekly (Starter) or Weekly (Growth+)
- Analysis depth: Post-purchase surveys, support analysis, inventory strategy, KPI planning, quarterly planning (varies by tier)
Bonus (Growth+): Comprehensive email marketing audit from specialist partners
Flexible plans give you complete control over costs. You pay for the essential CRO work - strategy, hypothesis generation, analysis, A/B test and project management - whilst design, development, and QA are billed separately at $70/hourly only when you need them.
This is perfect if you have an in-house design or development team, or if you want to manage exactly what gets built and when. You're not locked into paying for services you don't need.
Scale plans include everything - strategy, analysis, design, development, QA, and implementation - in one predictable monthly retainer. No surprises, no separate invoices, just complete care-driven CRO delivered autonomously.
Choose Flexible if: You have internal resources or want precise cost control
Choose Scale if: You want fully autonomous, hands-off CRO with everything included
Transparent pricing based on your monthly traffic.
We charge based on traffic volume because testing capacity and statistical significance directly correlate with session count. The more traffic you have, the faster we can run tests and deliver results.
Pricing:
- Starter (50K-75K sessions): $1,650/mo - 2 tests
- Growth (75K-150K sessions): $3,500/mo - 4 tests
- Scale (150K-350K sessions): $6,600/mo - 6 tests
- Enterprise (350K+ sessions): $10,700/mo - 8 tests
No long-term contracts. Cancel anytime.
Every plan includes our 30-day profitability guarantee.
Not sure which plan fits?
Book a discovery call - I'll help you find the perfect match for your business.
Our battle-tested frameworks and systems validate every hypothesis before we build.
Phase 1: Onboarding (First 5 days)
- Deep-dive into your business, customers, and psychology
- Comprehensive technical audit
- 25+ care-driven optimisation hypotheses
- Custom roadmap delivered
Phase 2: Operational (Continuous)
- Validate hypotheses through AI-trained buyer personas
- Ask: "Does this genuinely serve customer needs - not manipulate?"
- Design, develop, and implement winning tests
- Rigorous QA across all devices
- Launch and monitor
Phase 3: Ongoing Analysis (Monthly)
- Behavioural segmentation & data analysis
- Post-purchase survey analysis (Growth+ plans)
- Support ticket insights analysis (Growth+ plans)
- Inventory strategy (Growth+ plans)
- Monthly KPI planning (Growth+ plans)
- Quarterly strategic planning (Scale+ plans)
Yes - but as an addition to our battle-tested frameworks, not the foundation.
We've built a proprietary AI system that validates every hypothesis against your actual buyer personas before we build anything. This ensures we only create optimisations your customers will genuinely respond to.
How it works:
- Our frameworks identify conversion opportunities
- We generate psychology-first hypotheses
- AI-trained buyer personas validate each hypothesis
- We ask: "Does this genuinely serve customer needs—not manipulate?"
- Only validated hypotheses get built and tested
This approach achieves 84% test success rate vs 45% industry average - because we validate with your actual customers before building, not after.
AI enhances our care-driven methodology. It doesn't replace genuine customer understanding.
Simply upgrade to the next tier for more included tests and enhanced ongoing analysis.
We're completely flexible - scale up or down based on your business needs. No penalties, no long-term lock-ins.
Want to discuss expanding your plan? Your dedicated CRO manager can adjust your package anytime.
Yes. No long-term contracts. Cancel anytime.
We earn your business every single month through results - not by trapping you in contracts.
If we don't make you profitable within 30 days, you pay nothing more until we deliver. That's our guarantee.
Most clients stay because care-driven CRO compounds month after month - each winning test keeps generating revenue whilst new tests add even more. But you're never locked in.
We're confident our results will speak for themselves.
Zero micromanagement required. We operate completely autonomously.
We're an extension of your business - making decisions with your profit margins AND mission in mind, not billable hours.
Your involvement:
- Initial onboarding: 2-3 hours (interviews, strategy alignment)
- Weekly/bi-weekly meetings: 30-60 minutes (strategy updates, results review)
- Ad-hoc questions: Slack chat for quick questions
We handle everything else:
- Hypothesis generation
- Design and copywriting
- Development and implementation
- QA across all devices
- A/B test management
- Data analysis and reporting
You focus on running your business. We focus on adding $50K+ monthly to your revenue.
That's the partnership.
We integrate with your existing tools—no forced changes.
Analytics: Shopify Analytics, Microsoft Clarity, GA4
Testing: Intelligems
Management: ClickUp, Figma, Slack
Your data stays in your systems. We integrate seamlessly.
We sign NDAs before any work begins. Your data is protected - always.
Security measures:
- Non-Disclosure Agreement (NDA) signed upfront
- Limited access permissions (only what's necessary)
- Data stored in your systems (we don't migrate your data)
- Team access restricted to assigned personnel only
- Regular security audits
We treat your business like our own - that includes protecting your data like it's our own.
You maintain full control over all access permissions and can revoke them anytime.
Guaranteed profitability in 30 days. $50K+ monthly revenue boost within 60 days.
Tangible outcomes:
But more than numbers - you'll understand your customers deeply, remove friction authentically, and build genuine relationships that compound revenue month after month.
- Increased conversion rates (50-100%+ improvements common)
- Higher average order values
- Improved ROAS (return on ad spend)
- Enhanced customer lifetime value
- Sustainable, compounding revenue growth
Our 84% hypothesis success rate means tests consistently work.
Real client results:
- ForKeeps Merch: $2.3M added revenue (+70% conversion rate)
- Organic Muscle: 128% conversion rate increase
- CKitchen: $1.1M added revenue over 22 months
- Mayven Studios: 50% conversion increase in 2 months
For as long as care-driven CRO continues delivering massive ROI - which typically compounds over 6+ months.
Why long-term partnerships work:
- Each winning test keeps generating revenue permanently
- New tests stack on top of previous wins
- Deeper customer understanding leads to better hypotheses
- Compounding effects multiply over time
Typical timeline:
- Months 1-3: Foundation + initial wins ($50K+ monthly added)
- Months 4-6: Compounding effects visible (wins multiply)
- Months 7-12: Sustainable growth system established
- 12+ months: Category-leading conversion rates achieved
Most clients stay 12-24+ months because results compound. But there's no lock-in - cancel anytime.
We earn your business every month through genuine results, not contracts.
Three simple steps:
Step 1: Book a Discovery Call 30-minute conversation to discuss your traffic, goals, and biggest challenges. We'll explore if we're a good fit and map out your path to $50K+ monthly revenue growth.
Step 2: Get Your Free Audit We'll conduct a comprehensive CRO audit of your website, deliver 25+ psychology-first hypotheses, and show you exactly where your biggest revenue opportunities are.
Step 3: Choose Your Plan & Launch Select the plan that fits your traffic and business needs. We'll onboard you within 5 days and have your first A/B test live within 10 days.
Ready to grow with care-driven CRO?
Or have more questions? Email us: garyk@weblics.agency
Spending more on
Getting less back?
Your ads aren't broken. The store they land on is.
Send us your URL and we'll show you which leaks are eating your ad spend, and which one to fix first.
Just honest findings on your store, within 48h.