Website design that helps people buy

Updated in 2026: The original draft treated visual techniques as conversion tactics. This version starts with the customer’s decision and the evidence available on the site.

A redesign improves sales only when it improves a measured decision path. Define the intended action, find where suitable visitors hesitate or fail, and test whether the new design changes that behaviour.

Start with the decision

Name the audience, the action and the information they need before taking it. For a complex service, that may include fit, process, evidence, risk and a conversation. For a product, it may include compatibility, delivery, returns and total price.

Ask a representative visitor to explain what the company offers, who it serves and what they should do next. If they cannot, the page is unclear. This is a diagnostic exercise, not a universal five-second rule.

Replace reassurance with evidence

Generic claims such as “market-leading” and “customer-focused” are easy to add and impossible to assess. Show relevant work, named methods, product details, limitations, current policies and specific customer outcomes you have permission to publish.

A testimonial saying “great team” supplies sentiment. A case explaining the task, starting point, change and timeframe supplies evidence. Do not invent precision or present one result as a guaranteed benchmark.

Remove friction you can observe

Watch people use the key journey and inspect analytics, form errors, support questions and abandonment. Fix hidden costs, ambiguous buttons, forced account creation, broken validation and unnecessary fields before adding persuasion devices.

Slow or unstable pages can interrupt the journey, especially on mobile. Measure real-user performance alongside conversion and abandonment rather than assuming one loading threshold predicts sales. Google’s Core Web Vitals provide useful experience metrics, not a substitute for business data.

Test the change

Record the baseline, release the smallest coherent improvement and compare qualified outcomes over a sensible period. Note traffic mix, seasonality and concurrent campaigns. If the evidence is inconclusive, say so.

The useful design question is not “does this look modern?” It is “does this help the right person decide with less confusion and no loss of trust?”