Ecommerce Best Practices for Shopify Stores: A Sourced, Prioritised Guide

Most "best practice" lists are opinions with no source. This one is different: every practice below points to a page we opened and, from Shopify's developer docs, Google, the W3C, two government sites and the Baymard Institute's free research. We kept the list short, and we put a checklist first so you know what to fix before you read the detail.

The plain answer: for most small Shopify stores the biggest gains come from five things. Show the total cost and the return policy before checkout. Keep the page fast. Add as little as you can to checkout. Make email and returns easy. Make the store usable with a keyboard and a screen reader.

How to read the numbers. Many figures come from Baymard, which tests large retail sites and surveys US shoppers. They show what real shoppers struggle with, not a promise for your store. We say where a figure is old or narrow. Test changes on your own store before you trust any of them.

What to fix first: a prioritised checklist

We ordered this by two things: how strong the evidence is, and how little work it takes. The ranking is our judgement. The facts behind each row are in the sections below.

If you can only do three things this month, do the three priority 1 rows.

Speed: pass the three Core Web Vitals first

Google's Core Web Vitals have three targets: Largest Contentful Paint (LCP) within 2.5 seconds, Interaction to Next Paint (INP) of 200 milliseconds or less, and Cumulative Layout Shift (CLS) of 0.1 or less. web.dev says to judge them at the 75th percentile of real page loads, split by mobile and desktop. Google's Search Central page recommends good scores for success in Search (web.dev on Core Web Vitals, Google Search Central).

Shopify's own theme performance docs put a short list under "Essential practices", described as having the largest impact across the broadest range of themes. The ones a store owner can act on:

• Never lazy-load the image that is the LCP element. Load images visible at the top of the page eagerly and give the main image fetchpriority="high".
• Do not hide the main image behind a fade-in animation.
• Keep product info, hero content and navigation in Liquid and HTML, not in client-side JavaScript.
• Remove or defer app scripts that block rendering, and audit scripts that are not earning their cost. Shopify names too much JavaScript as a main cause of poor INP.

One more line from the same docs is worth knowing: Shopify warns about "fake performance apps" that cheat Lighthouse and PageSpeed tests, for example by serving different pages to crawlers. A better score from an app like that does not help your shoppers. Measure with real-user data where you can, not only a lab score.

Where to go deeper: our theme performance guide goes through the theme side step by step. We did not run any speed tests for this post, so we quote no "before and after" numbers.

Product pages: answer the questions before they are asked

The product page is where nearly every shopper decides, so small gaps cost sales. Baymard's current product page article says up to 62% of the leading sites it benchmarks have mediocre or worse product page UX. These are the practices from it that we think matter most for a small DTC store.

Show the total cost near the buy button. Baymard found 67% of benchmarked sites do not give a total order cost estimate on the product page. Shipping, tax and fees that appear for the first time at checkout surprise shoppers, and some leave. For free-shipping rules, see our free shipping guide.

Link the return policy from the product page. In Baymard's survey, 60% of respondents look for return information on the product page, and 15% said they had abandoned an order in the last quarter because the return policy was unsatisfactory. Yet 44% of benchmarked sites did not display or link to it there.

Use buttons for sizes and show scale. Baymard saw drop-down size menus hide which sizes are in stock. It also found 42% of shoppers try to judge a product's size from its images, so one photo with an object of known size helps.

Do not leave sold-out shoppers at a dead end. Baymard found 30% of shoppers are likely to abandon when told a product is out of stock. Its advice is to let people order temporarily out of stock products where you can really supply them. Whether that fits your stock is your call. Our pre-order apps guide covers the Shopify side.

Write descriptions people can scan. Baymard says every tester relied on the description at some point, and that descriptions structured by feature highlights drew more in-depth reading than plain text blocks.

Put product data in markup Google can read. Google's documentation says merchant listing structured data can show price, availability, review ratings, shipping and return information in search. Required properties include the product name, an image and an offer. Shopify's Liquid structured_data filter outputs a schema.org Product for a product object, which many themes already use, so check what yours emits rather than adding a second copy. Validate with Google's Rich Results Test. Google also documents return policy and shipping policy markup, if you want those details to appear.

Cart and checkout: remove friction, add little

Baymard puts the average documented cart abandonment rate at about 70% (70.19% on its research article, 70.22% on its list of 50 studies). It also says a large share of that is shoppers who are only browsing, which no design change fixes. Of the rest, 17% of surveyed US shoppers said they had abandoned an order in the last quarter because checkout was too long or complicated.

On Shopify you do not build the checkout page from scratch. Shopify's developer docs describe checkout as Shopify's, extended through apps and extensions, and say those options are upgrade-safe so merchants can take new features such as Shop Pay and one-page checkout. So your levers are narrower than on a custom site:

• Add less. Every extension, upsell block or extra field you add works against what Baymard measured. Baymard's 2024 article says most sites need about 8 form fields in checkout, while the average it found was 11.3, and that the number of fields matters more than the number of steps.
• Do not push accounts early. Baymard says 42% of benchmarked sites ask for account creation before or during checkout, and that this distracts some shoppers into leaving. Its advice: guest checkout always, and offer an optional account on the confirmation step.
• Use what already reduces typing. Baymard recommends a single name field and an address lookup feature. Check whether your payment and address settings already support these.
• Be careful with trust badges. Baymard found 19% of shoppers had abandoned a checkout in the last quarter because they did not trust the site with their card details. In its tests any visual security cue raised the sense of safety, and well-known third-party seals tested best, while a badge you make yourself is only your own claim. Do not add a pile of badges to look safe.

We could not open Shopify's Help Center pages today (they blocked our fetch tool), so we do not name the exact admin settings for guest checkout or customer accounts. Check those in your admin.

Trust and policies: say it plainly, early

Trust here means two things: shoppers can see what happens if something goes wrong, and you meet the law where you sell.

• State delivery and returns in plain words on a page and on the product page, as above.
• UK sellers and sellers into the UK: the UK government's distance selling guidance lists what you must tell customers before they order, including your business name, address and contact details, the full price with taxes, delivery costs and times, and how to cancel. It says customers can cancel up to 14 days after delivery, and that if you do not tell them, the right lasts up to 12 months (GOV.UK).
• EU buyers: the EU's consumer site describes a 14-day right to withdraw from an online purchase without giving a reason, counted from delivery for goods (Your Europe).
• US: we did not open a US source today, so we say nothing about US rules here. Ask your lawyer.
• Reviews: Baymard reports that 95% of tested users relied on reviews, and relied on the ratings distribution summary even more than on individual reviews. It also says 43% of benchmarked sites lacked one, and that shoppers often read the negative reviews, so a reply from the store helps. Our review apps guide compares tools.

This is general information, not legal advice. Laws differ by country and product type, and some goods are exempt.

Email and retention: fewer, better, easy to leave

Baymard's August 2026 survey of 1,083 US shoppers found 48% had unsubscribed because emails came too often from one retailer, 46% because of too many newsletters overall, and 44% because they lost interest in the content. Baymard's advice is to give people a preference centre that controls how often they hear from you, instead of an all-or-nothing unsubscribe.

Two technical rules matter as well. Google's sender guidelines say all senders to Gmail need SPF or DKIM, TLS and a spam rate below 0.3% in Postmaster Tools. Bulk senders (more than 5,000 messages a day to Gmail) also need SPF, DKIM and DMARC, and marketing messages must support one-click unsubscribe plus a visible unsubscribe link (Google). Most email platforms handle the headers if you set up your sending domain, but you should check that you did.

For retention beyond email, Baymard's survey says 51% of shoppers rank order status or delivery tracking among the most important account features, and 24% rank loyalty programmes as a high priority. Its order-tracking article lists six items shoppers want: expected delivery date, a status progress bar, carrier name, a linked tracking number, shipping history and a summary of the package contents. It found 67% of test sites did not consistently show all of them.

We did not open a source on abandoned-cart email results today, so we do not quote any. One thing Baymard does tell us: 42% of shoppers who abandoned a cart in the last three months said they were just browsing, so do not expect a recovery flow to win most of them back.

Analytics: count the steps, respect consent

You need to see where shoppers drop out. Google Analytics 4 documents a standard set of ecommerce events, including view_item_list, view_item, add_to_cart, view_cart, begin_checkout, add_shipping_info, add_payment_info, purchase and refund. Its guide recommends enabling debug mode while you set up and setting currency on value data (Google).

On Shopify, checkout tracking runs through Shopify's customer events and Web Pixels API, where app pixels load in a sandbox (Shopify). Shopify's Customer Privacy API is the route for checking and applying a visitor's consent to pixels and checkout (Shopify). If you sell in the UK or EU, make sure your cookie banner really connects to it.

Two habits help: compare your tool's purchase count with Shopify's order count each month, and read field Core Web Vitals in Search Console rather than only a lab score.

Returns: plan for them before launch

Baymard's August 2026 survey found 52% of US shoppers returned an item in the past 12 months, and that frequent returning is falling (28% returned more than once, down from 44% in 2024). Baymard suggests two possible reasons and says it cannot tell which: better shopping tools, or shoppers keeping wrong items to avoid strict windows and fees. Its advice either way is to make the policy highly visible and easy to understand before checkout.

An older Baymard study (2019) found 54% of sites had significant usability problems in their returns flow, and that returns caused the most trouble in its account testing. Treat that as dated, but the direction is clear: test your own return process as a customer would, from the order email to the label.

Accessibility: aim for WCAG 2.2 AA and test with a keyboard

WCAG 2.2 became a W3C Recommendation on 5 October 2023. For stores, several of its newer criteria hit checkout and forms directly:

• Target size (2.5.8, AA): pointer targets at least 24 by 24 CSS pixels, or enough spacing.
• Focus not obscured (2.4.11, AA): a sticky banner must not completely hide the item that has keyboard focus.
• Redundant entry (3.3.7, A): do not make people retype information they already gave in the same process.
• Accessible authentication (3.3.8, AA): do not require remembering a password or solving a puzzle as the only way to log in.
• Contrast (1.4.3, AA): text needs a contrast ratio of at least 4.5:1, with 3:1 for large text.

Shopify's theme accessibility notes add store-specific items: visible keyboard focus, a skip link, zoom not disabled, descriptive alt text on product images, labels on all form fields, and price or availability changes announced to screen readers when a variant changes.

The W3C's Easy Checks page gives a quick first review: alt text, page titles, headings, contrast, skip link, visible focus, page language and zoom. It says these checks are quick, not exhaustive. Baymard's 2021 benchmark of 33 large ecommerce sites found 94% of them incompliant with WCAG 2.1 AA across four basic areas (images, links, form fields and keyboard navigation). That figure is old and from big sites, but it shows how common the gaps are.

On law: the European Commission's page on the European Accessibility Act lists e-commerce among the covered services. It does not give details we could rely on for your case, so ask a specialist where you sell. For the build side, see our theme accessibility guide.

What we left out, and why

Lists like this one usually include countdown timers, exit popups and "scarcity" banners. We did not include them because we found no source we opened today that supports them. If you test them, test them on your own store, and keep the claims true.

Before you launch or redesign

If you are launching, run through our Shopify launch checklist next. If you want help finding which of the rows above cost your store the most, talk to us. We would start with your analytics and a checkout walk-through on a phone.

Frequently asked questions

What are the most important ecommerce best practices?

Show total cost and the return policy before checkout, keep pages fast, keep checkout short, make email and returns easy to manage, and make the store accessible. The checklist above puts them in order.

What should a new Shopify store fix first?

The three priority 1 rows: costs and returns on the product page, the main image and app scripts for speed, and guest checkout.

Do trust badges increase sales?

We found no proof that a badge raises sales. Baymard's tests say visual security cues raise the sense of safety and that known third-party seals did best, while self-made badges are only your own claim. Use few, and only true ones.

Does page speed matter for ecommerce?

Google recommends good Core Web Vitals for Search success, and Shopify says performance affects conversion, repeat business and search rankings. We did not measure a sales effect for this post, so we do not quote one.

Should I force customers to create an account?

No. Baymard saw early account prompts distract shoppers. Offer an optional account after the order is placed.

How many fields should checkout have?

Baymard says most sites need about 8 form fields, against an average of 11.3 in its 2024 benchmark. On Shopify, the base checkout is Shopify's, so your job is not to add fields.

Is my store legally required to be accessible?

It depends on where you sell. The EU lists e-commerce as a covered service in its Accessibility Act. Ask a specialist about your countries.

Which analytics events should I track?

At least product views, add to cart, checkout start and purchase, plus refunds. GA4 documents standard names for each.