Chris Raad · 20 November 2026

Ecommerce UX best practices that research has measured, by stage of the shopping journey, with the source for each and a way to check your own store in minutes.

The ecommerce UX best practices worth your time are the ones somebody measured: show the total cost and a delivery date before checkout, make search and filters work on a phone, put the answers to product questions next to the buy button, keep the checkout short, and keep pages fast. Contentsquare's 2026 benchmark found phones bring 69.9% of traffic while desktop converts 74% higher, so most of the gap sits on the small screen.

The "best practice" lists that rank for this question mostly repeat each other without sources. This one groups practices by stage of the shopping journey, names the study behind each, marks the ones that matter most on mobile and says how to check your store. Where the evidence is weak or comes from a vendor, we say so.

What counts as an ecommerce UX best practice?

A practice counts when a usability study, a large benchmark or a controlled test shows that it changes what shoppers do, and when you can check your own store for it in minutes. Opinions about what looks modern do not count, and neither do lifts quoted without a sample size.

The sources behind this guide, and what each can and cannot tell you:

SourceWhat it measuredLimits
Baymard InstituteModerated usability testing plus benchmarks of 150 to 344 leading US and European sites, scored against its guidelinesLarge retailers, not Australian stores; Baymard sells audits and research
Nielsen Norman GroupUsability studies of ecommerce sites over two decadesSome articles are a decade old; the behaviour has held
Contentsquare 2026 benchmark99 billion sessions on 6,500+ sites, Q4 2025 against Q4 2024Global; Contentsquare sells the analytics
Shopify store speed studyCore Web Vitals against conversion across actively selling Shopify stores, early 2026Correlation, and Shopify says so
Australia Post eCommerce Report 2026Omnibus survey of at least 1,500 Australians, November 2025Stated preferences, not observed behaviour

Baymard benchmarks the biggest sites, with the largest UX budgets, and expects smaller sites to do worse.

Which ecommerce UX practices matter most?

The practices that remove a reason to stop buying come first: unknown total cost, unknown delivery date, a product people cannot find, a question the product page does not answer, and a slow or broken page. This table ranks them by stage, with the evidence and a check you can run today.

StagePracticeEvidenceMobile priorityCheck your store
NavigationProduct categories as the first level of the mobile menu22% of benchmarked sites don't (Baymard 2026)HighOpen your menu on a phone: do product types appear first?
SearchSearch handles product types, features and misspellings56% of sites fail to support search adequately (Baymard 2026)HighSearch for a product by its feature ("waterproof"), a misspelling and "returns"
Category pagesColour variations combined into one list item42% of sites don't (Baymard 2025)HighDoes one shirt in four colours appear once or four times?
Product pageShipping cost or total cost shown near the buy button67% of sites don't (Baymard 2026)HighCan you see delivery cost without adding to cart?
Product pageSize and variant options shown as buttons, not drop-downs57% of sites don't (Baymard 2026)HighCan you see which sizes are sold out before tapping?
Product pageReturn policy linked from the product page44% of sites don't; 60% of users look for it there (Baymard)MediumFind your returns terms from a product page in one tap
CartDelivery date, not delivery speed48% of sites don't (Baymard checkout 2026)HighDoes your cart say "arrives Thursday" or "3 to 5 business days"?
CheckoutGuest checkout, account offered after purchase18% of US shoppers abandoned over forced accounts (Baymard)HighCan you buy without creating an account?
CheckoutFewer form fieldsAverage checkout shows 23.48 form elements; 12 to 14 is achievable (Baymard)HighCount the fields between cart and payment
CheckoutSpecific error messages93% of mobile sites don't adapt them (Baymard 2026)HighEnter a short phone number: does the error say what is wrong?
SpeedLCP under 2.5 seconds on phonesAbout 3.5% lower conversion per 100 ms slower (Shopify 2026, correlation); Vodafone A/B testHighPageSpeed Insights field data for your product page
TrustDelivery options and timing stated early70% of Australians less likely to buy with poor delivery communication (Australia Post 2026)HighIs delivery timing on the product page as well as checkout?

How should navigation and search work on an online store?

Navigation should put product categories first and use filters for attributes, and search should return the right products for the words shoppers use. Baymard's testing found roughly half of participants prefer search to find products and half use the menu, so both paths carry sales.

Baymard's 2026 benchmark rates 61% of desktop and 65% of mobile sites "mediocre" to "poor" on homepage and category navigation. The findings that apply to most Shopify stores:

  • Make product categories the first level of the mobile menu. Where they sat behind a single "Shop" item, some test participants closed the menu and turned to search. 22% of benchmarked sites don't.
  • Use filters, not separate categories, for attributes that products share. 79% of sites split products such as "skinny jeans" and "slim jeans" into categories, which stops shoppers combining criteria.
  • Keep each menu level to about 10 options. Most participants felt overwhelmed past that.

For search, Baymard's 2026 search benchmark found 46% of desktop, 58% of mobile and 64% of app sites "mediocre or worse". The weakest query types were non-product searches (66% of sites have issues), where people type "returns" or "shipping"; in Baymard's account testing, 34% of participants used the search field for that kind of content. On mobile, put a visible search button next to the field: 27% of mobile sites don't, and some testers never noticed the keyboard's own button.

Check your store: search for a product by a feature, by a common misspelling and by "returns", on a phone. If any of the three returns nothing useful, that is your first fix.

What makes a good category page?

A good category page shows each product once, with enough images and detail to choose which to open. Baymard's 2025 benchmark rates 78% of mobile sites "poor" to "mediocre" on product lists, against 58% on desktop.

Two fixes cover most Shopify collections:

  • Combine colour and size variations into one list item with swatches. 42% of sites show each colour as a separate product, which clutters the list and hides the variation a shopper wants.
  • Give each list item more than one image. 80% of sites don't offer three or more thumbnails, so shoppers open product pages one by one to see a second angle.

What belongs on an ecommerce product page?

A product page has to answer what the product is, whether it fits, what it costs delivered, when it arrives and what happens if it is wrong. Nielsen Norman Group attributed 20% of task failures in its ecommerce studies to incomplete or unclear product information, and Contentsquare found 1 in 3 visits now starts on a product page.

Baymard's 2026 product page benchmark found only 48% of desktop and 38% of mobile sites reach "decent" or better. The practices with the widest gaps:

PracticeSites that don't (Baymard)What testing showed
Show a total cost estimate near the buy section67%Without it, shoppers add to cart only to find the delivery cost
Use buttons for size selection57%Drop-downs hid sold-out sizes until the shopper opened the menu
Link the return policy from the product content44%60% of surveyed users look for return terms on the product page
Provide an in-scale image37%42% of testers tried to judge size from the images
Respond to negative reviews89%Testers checked negative reviews often and took replies as a sign of good service

Reviews carry more weight than most of the page. In Baymard's earlier product page testing, 95% of users relied on reviews, and a ratings distribution summary was the most used part of the reviews section.

What to fix first on your own product page, in order, is its own guide:

How do you reduce friction in the cart and checkout?

Show every cost before checkout, give a delivery date, let people buy as guests and cut form fields. Baymard's survey of US shoppers who abandoned a checkout, excluding those who were only browsing, ranks extra costs first at 40%, then slow delivery (20%), not trusting the site with card details (19%), forced account creation (18%) and a long or complicated checkout (17%).

Baymard's 2026 checkout benchmark rates 59% of desktop and mobile checkouts "mediocre" or worse. The fixes that apply to most stores:

  • Delivery dates instead of delivery speeds. 48% of sites show "3 to 5 business days"; testers stopped to count days on a calendar.
  • A cutoff countdown ("order in the next 43 minutes for delivery Tuesday") instead of a static cutoff time. 83% of sites don't.
  • Quantity changes with plus and minus buttons in the cart. 95% of sites don't use buttons or buttons with a text field, and testers placed orders with the wrong quantity.
  • Fewer fields. The average US checkout shows 23.48 form elements by default; Baymard puts an achievable flow at 12 to 14.

In Australia, Shopify research cited in the Australia Post 2026 report found 52% of Australians had abandoned a cart because checkout was too long or complicated, rising to 67% for higher spenders.

On Shopify, the cart is part of your theme and open to change on every plan. The checkout is not: apps that change the information, shipping and payment steps are available only on Shopify Plus. Most of the fixes above therefore belong in the cart and on the product page. Our guide to what you can change in Shopify checkout covers the limits plan by plan, and cart abandonment rate covers how to measure the problem first.

Which UX practices matter most on mobile?

On a phone, the practices that save typing, scrolling and guessing matter most, because mobile visits are shorter and see less of each page. Contentsquare measured 2 minutes 20 seconds per mobile session against 4 minutes 46 on desktop, and a scroll rate of 45.2% against 50.5%.

Baymard's 2026 mobile benchmark of 150+ sites found only one rated "good". The mobile findings with the widest gaps:

  • Specific validation errors. 93% of mobile sites show the same generic message for every error, such as "invalid phone number" when the number is incomplete; some testers could not continue at all.
  • Address lookup or validation. 54% of sites don't, and typing errors on a touch keyboard lead to failed deliveries.
  • A "view all" option at each level of the category menu. 42% of mobile sites don't, which traps shoppers in narrow subcategories.

A sticky add to cart button is a common mobile recommendation, and its design decides the result. GrowthRock, a US CRO agency, published a mobile product page test on a supplement store with more than 3,000 orders per version over 14 days. A sticky button that opened a slide-up drawer with flavour and quantity raised orders 5.2% at 98% significance. A sticky bar that only scrolled shoppers back to the buy section lifted add to cart clicks about 6% but showed no significant change in orders. Baymard adds that a sticky button should keep white space around it rather than span the full width of the screen (guideline 791).

Part of the mobile gap is not friction: Contentsquare reads it as people starting on a phone and finishing on a computer. In recordings, rage taps, dead taps and repeated scrolling past the buy button are friction; a calm read of a product page that ends without a struggle may be research.

How much does page speed matter for conversion?

Speed matters most for load time and responsiveness on phones. Shopify's study of its own stores found conversion tends to be about 3.5% lower for every 100 ms slower the main content loads, and about 1.5% lower for every 32 ms slower a store responds to taps. Stores with a 2.5 second LCP reported roughly 30% lower conversion than stores at 1.5 seconds.

Shopify calls this a correlation, not proof of cause, and found mature, high-volume stores show roughly a third of the effect. Layout shift (CLS) showed no such pattern in its data. The strongest controlled evidence is older: Vodafone ran an A/B test in which the only difference was a 31% better LCP, and the faster page made 8% more sales.

Fewer than half of mobile sites pass all three Core Web Vitals: 48% in 2025, on HTTP Archive's measurements of millions of sites. Shopify's own advice is to keep your theme updated and audit apps, because third-party apps add JavaScript that slows a store. Our guide to speeding up a slow website covers the fixes, and our Core Web Vitals research covers what Google measures.

Check your store: run a product page, not the homepage, through PageSpeed Insights and read the field data for phones. LCP over 2.5 seconds or INP over 200 ms is a problem real customers are having.

What do Australian shoppers need to see about delivery and trust?

Australian shoppers want delivery timing, options and cost stated early and clearly. In Australia Post's November 2025 survey of at least 1,500 consumers, 70% said poor delivery communication at checkout makes them less likely to complete a purchase, and 69% want a range of delivery options at checkout, including parcel lockers.

The same survey found 26% expect same-day or next-day delivery when an order is time-sensitive, and 43% would pay extra for it. Australia Post's advice is to show delivery options early, on product pages, and again at checkout. That matches Baymard's earlier product page testing, in which 64% of users looked for shipping costs on the product page before deciding to add to cart.

Delivery detail can be the conversion problem on its own. The Flower Room, a Newtown florist we rebuilt, had a delivery suburb picker that showed one suburb per postcode and hid places such as Enmore and St Peters. After our September 2026 fix it lists 791 suburbs, up from 213 (case study).

Trust at the payment step is the other half. 19% of US shoppers in Baymard's survey abandoned because they did not trust the site with card details, and 13% over the returns policy. Familiar payment options help here; our guide to Afterpay, Zip and Shop Pay covers what the Australian evidence shows.

How do you check your own store against these practices?

You can check a store against this list in an afternoon with free tools; reading the evidence well takes days. Let your own data pick the first fix, because the table above ranks what is common, not what is broken on your store.

  1. Install Microsoft Clarity (free) and filter recordings to phone sessions that reached the cart and did not buy. Clarity flags rage clicks, dead clicks, excessive scrolling and quick backs for you.
  2. Read the GA4 purchase funnel split by device. A step where mobile drops far more than desktop is where to start; our guide to GA4 on Shopify covers why the numbers may not match Shopify's.
  3. Buy something on your own phone, over mobile data, as a new customer. Note every point where you had to guess.
  4. Run your best-selling product page through PageSpeed Insights and read the phone field data.
  5. Fix the biggest problem you found, and measure revenue per visitor against the same weeks before the change and a year earlier.

Installing Clarity takes minutes; watching 200 recordings and turning them into ranked fixes is where the time goes. Our one-hour diagnosis walks through these steps in more detail.

Which best practices should you be sceptical of?

Be sceptical of any practice quoted with a lift but no sample size, test length or metric. Small tests produce large numbers by chance: Optimizely's analysis of 127,000+ experiments on its platform found about 12% won on their primary metric (2023), so a list of only winners leaves out the losers.

Before you adopt a practice from any list, this one included:

  • Check who was studied. Baymard tests large US and European retailers. An Australian florist shares most of their usability problems, but not their order values or delivery promises.
  • Check the implementation, not the label. The sticky add to cart result above won in one design and did nothing in another.
  • Check your own data before building. A practice that 67% of sites miss may already be right on yours.

Most stores also cannot A/B test small changes: detecting a 10% relative lift from a 2% conversion rate takes about 161,000 sessions (CRO cost guide). Below that traffic, ship the fixes the evidence supports, measure them against the baseline, and save tests for large changes on busy pages (A/B testing on Shopify). If the problems run deeper than any list of fixes, our guide on whether to redesign or optimise covers when a rebuild is the right call.

Sources

Questions

What are the most important ecommerce UX best practices?

The practices with the strongest evidence are the ones that answer cost and delivery questions before checkout, help people find products on a phone, and keep pages fast. Baymard's survey of US shoppers puts extra costs first among reasons for abandoning a checkout (40%), and Australia Post's 2026 report found 70% of Australians are less likely to buy when delivery communication at checkout is poor. Start there, then work through search, product pages and the checkout form.

Why does mobile convert worse than desktop?

Contentsquare's 2026 benchmark of 99 billion sessions found desktop converts at 3.4% and mobile at 2.0%, so desktop converts 74% higher, while phones bring 69.9% of the traffic. Mobile visits are also shorter (2 minutes 20 seconds against 4 minutes 46) and see less of each page. Some of the gap is people researching on a phone and buying later on a computer; the rest is friction you can find in session recordings.

Do Baymard's findings apply to small Australian Shopify stores?

Baymard benchmarks large US and European sites, so its percentages describe big retailers, and Baymard itself says smaller sites are likely to perform worse. The usability findings come from watching people shop, so the behaviour transfers. Check each practice against your own recordings and funnel before you build anything.

Is a sticky add to cart button worth adding on mobile?

Sometimes, and the design decides it. In a published test with more than 3,000 orders per version, a sticky button that opened a slide-up buying drawer raised orders 5.2%, while a sticky bar that only scrolled people back to the buy area showed no significant change. Long product pages gain the most; short pages where the button rarely leaves the screen gain little.

How do I check my store's UX without paying for an audit?

Install Microsoft Clarity (free) and filter recordings to phone sessions that reached the cart and left. Check the GA4 purchase funnel by device, buy something on your own phone over 4G, and read your Core Web Vitals in PageSpeed Insights. That takes an afternoon to set up; reading a few hundred recordings well takes days.

We'll show you what your customers do on it.

Apply for a free teardown