DTC technology × growth systems — built to compound.
DTCUP / FIELD NOTES / BUILD

Website UX Audit: Find What Stops Customers from Buying

A practical way to investigate a difficult buying journey, separate website problems from offer problems, and choose the improvements worth making first.

Published Mar 29, 2026Updated Sep 20, 20267 min read
Jack Hu
Jack Hu

7 years of experience building and operating brand websites, specializing in website development, paid advertising and SEO. Built a brand website from launch to US$10 million in annual sales.

The Short Answer

I would start with one important buying journey and the customer questions it needs to answer. Check the offer and traffic before assuming the design is at fault. Then reproduce problems on real devices and compare what you see with support questions, customer observations and relevant page data.

Fix verified purchase blockers first. Improve missing or confusing information where customers make the decision, and check that the business can deliver the promise. Leave the audit with a short list of actions, owners and follow-up checks. A long list of visual opinions is not a useful substitute for a clear decision.

Key takeaways

  • Investigate the offer, audience and stock situation before blaming the website.
  • Follow a real journey from the traffic source through product choice and checkout.
  • Prioritize reproducible failures and important unanswered customer questions.
  • Judge improvements alongside returns, support and commercial outcomes.

A website can look polished and still make buying feel difficult. A visitor may like the product but be unable to choose a size, understand the delivery cost or tell whether a bundle includes the part they need. Those are buying problems. A fresh color palette will not necessarily solve them.

When I review a store, I start with the question a customer is trying to answer at each step. Then I check whether the website answers it clearly and whether the business can deliver what it promises. This approach works on Shopify and other commerce platforms. Platform settings matter later, when we choose how to fix a specific problem.

The aim of an audit is to leave with a short, justified list of improvements. A document containing 80 observations but no order of importance is more work for the owner, not a useful result.

First separate a website problem from a business problem

Before opening the theme editor, write down what has changed. Did sales fall after a redesign? Did conversion drop when advertising expanded to a new audience? Are the products people want out of stock? Has delivery become more expensive? A website can only explain and support an offer; it cannot make every offer attractive.

I would compare the same kinds of visitors before blaming the page. Returning customers and people who have never heard of the brand arrive with different levels of trust. A broad advertising campaign can lower the overall purchase rate while still bringing in profitable new customers. Conversely, a promotion can improve conversion while leaving less money after costs.

Use a simple note with four lines: what changed, who was affected, where buying became harder, and what evidence supports that view. If the evidence is missing, say so. For example, “Mobile sales are down, but we have not checked whether mobile traffic became less relevant” is a better starting point than “Our mobile design is bad.”

Walk through one real buying journey

Choose an important product and the page most new visitors land on. Start from the actual advertisement, search result or email, rather than always entering through the homepage. Use a phone as well as a computer, and open a fresh browser session so that saved choices do not make the task artificially easy.

Follow the journey through five customer questions:

  1. Am I in the right place? The first screen should connect with the promise that brought the visitor here. An ad about a compact desk should not lead to a page where the customer must search for that desk again.
  2. Which option suits me? Show the differences that affect the decision: size, material, capacity, fit, compatibility or use. Replace internal product terminology with an explanation a new customer can use.
  3. Can I trust this choice? Provide genuine evidence, clear limitations and practical information. A close-up photograph, a measured dimension or a useful return explanation can answer a question that a slogan cannot.
  4. What will I pay and when will it arrive? Make the available shipping information and conditions easy to find. Do not promise a delivery date that operations cannot reliably support.
  5. Can I complete the order? Test options, quantity changes, cart updates, checkout entry and error messages. A visible button is not proof that the action behind it works.

For each problem, record the device, page address, exact action and observed result. “The size selector resets when I return from the cart on my phone” gives a developer something reproducible. “The page feels awkward” does not.

Find out whether customers share your difficulty

An owner’s familiarity with the catalog can hide confusing details. Ask a few people who resemble the intended customers to perform a realistic task without coaching. For example: “Find a jacket for wet-weather commuting and decide which size you would buy.” Ask them to explain what they are thinking. Do not tell them which menu or button to use.

These sessions reveal possible problems; a small group does not establish how common each problem is. Combine what you observe with support questions, returns reasons, on-site searches and page data. If visitors ask the same compatibility question in support and hesitate at the same point in a session, that is a stronger reason to improve the explanation.

For Shopify stores, the behavior reports include views of the buying path and on-site search. They can help locate where to investigate, but the numbers alone do not explain why someone left. Shopify’s behavior-report documentation describes those reports and their definitions.

Handle recordings and customer information under the site’s consent and privacy rules. The useful observation is the difficulty completing a task, not someone’s private checkout details.

Rank problems by consequence, reach and confidence

I use three questions to decide what deserves attention first. Does this stop or materially weaken a purchase? How many relevant visitors encounter it? How sure are we about the cause? Then I consider the work and risk involved in changing it.

Observation My starting priority Reason
An in-stock option cannot be added to the cart Fix promptly The buying task is broken and the failure can be reproduced.
Important product differences are missing Investigate and improve early Customers cannot confidently choose, especially if support confirms the confusion.
A delivery message is visible only after a long search Improve where the decision happens A relevant question is being answered too late.
A decorative animation feels dated Usually lower It may matter for brand presentation, but first establish what customer problem a replacement solves.

This is a decision aid, not a scoring formula that removes judgment. A problem affecting a smaller accessibility group can still prevent people from completing an essential task. Do not dismiss it simply because its measured reach is small.

Fix the explanation and the underlying cause together

Consider a hypothetical furniture store whose customers keep asking whether a table fits a small dining room. A useful change might combine a simple measurement drawing, photographs that show scale, a clear dimensions table and advice on measuring the available space. The change is not merely “add more content.” It is to make one decision easier.

Now check the dependencies. Are the dimensions correct for every option? Can the content team keep them current? Does the drawing remain readable on a phone? Can a keyboard user open the extra information? Is the delivery promise consistent with the shipping policy?

Forms deserve the same attention. Keep labels attached to their fields and explain an error beside the action that failed. W3C’s form-label guidance explains why visible and programmatically connected labels matter. A placeholder that disappears while someone types is a poor substitute for a clear label.

Decide what a useful result would look like before release

For a broken cart control, the first success condition is straightforward: the intended action works reliably on the affected devices. For unclear product information, improvement is less immediate. You might look for fewer repeated support questions, more completed selections and better purchase completion among comparable visitors. Also watch returns: persuading more people to buy the wrong product is not a good outcome.

Keep the period, audience and measurement definition consistent. Record promotions, price changes and advertising changes alongside the website release. A before-and-after increase is useful evidence, but it does not isolate the effect of design if several other things changed at once.

If traffic is limited, fix reproducible defects and improve obvious misunderstandings first. Do not promise a statistically reliable sales lift from a handful of orders. Treat uncertain improvements as hypotheses to revisit.

Leave the audit with three decisions

My preferred output is one page: what to fix now, what to investigate next, and what to leave alone for the moment. Each chosen improvement needs an owner, a clear description of the new behavior and a way to check it after release.

If the store works but feels hard to use, start with one important buying journey. If the core purchase path is broken, resolve that before visual refinements. If the problem is an unconvincing offer, address the offer with the commercial team instead of asking design to disguise it.

For the next step, use the 30-day improvement plan to schedule the work. If the team is proposing a full rebuild, first read how to choose between targeted fixes and a redesign.

Sources

  1. help.shopify.com — behaviour-reports
  2. www.w3.org — labels

Still have a question?Ask the DTCUP Community ↗

FAQ

Questions related to this insight.

Clear scope, practical constraints and direct answers for the decision ahead.

Do I need to redesign the homepage before doing an audit?

No. Start from the pages customers actually enter and follow one important buying journey. A homepage redesign may not address the difficulty you find.

What can I learn if my store has very little traffic?

You can verify broken controls, missing information and repeated customer misunderstandings. Small samples are not enough to claim a reliable sales lift.

Does this approach apply outside Shopify?

Yes. The customer questions and prioritization apply across commerce platforms. Reporting definitions and implementation steps depend on the platform you use.

Your workspace

Selected services

DTCUP / YOUR NEXT PERSPECTIVE

Stay curious.
Stay connected.

Ideas for the work behind growth—websites, brands and the apps that connect them.