UX method

Why your website is not converting: investigate before redesigning

A practical way to separate measurement, acquisition, offer and journey problems before committing to a UX redesign.

The button is not always the problem

You get visits but few enquiries. Changing a button colour sounds straightforward. Understanding what is happening takes more work. That final number combines very different situations: people who do not intend to buy, people who cannot understand the offer, and people who try but cannot complete the task.

Before redesigning a page, find where those situations diverge. This is a framework for an investigation, not a remote diagnosis of your website or a promise of a particular uplift.

1. Define what you are counting

A click on Submit is not an enquiry received. Check that the event represents a real confirmation and that one submission is not counted several times. For a sale, distinguish starting payment from an accepted order.

Choose the denominator too: visitors, sessions or people who started a journey. Keep the definition consistent across comparable periods. Look at absolute counts as well as rates. A handful of enquiries is not enough to explain every fluctuation.

2. Check the promise made before the visit

Someone looking for a definition has a different intention from someone comparing providers. Examine traffic sources and landing pages separately, then compare what was promised with what the page actually offers.

A fictional example: a post promises a free template but links to a quote request. Few enquiries would not prove the form is the problem. The promise and the destination do not match. Start there rather than with a fresh visual design.

3. Find the unanswered question

Before acting, people may need to understand scope, timing, cost or what happens next. For a shop, check fees, delivery and returns too. Baymard’s cart-abandonment research separates these concerns from difficulties with checkout itself.

That ecommerce research is not a conversion benchmark for your service business. Use it to raise questions, not to assign a cause to every departure. Review actual enquiries and the questions that repeatedly come up before a sale.

4. Try the task, not just the page

Choose a specific task on the relevant device: finding information, requesting a quote or buying an item. Follow it through to confirmation. Record the step, the observed obstacle and its effect, with an example or screenshot that contains no personal data.

Check links, fields, errors, going back, keyboard use and zoom. Where possible, observe people from the intended audience trying the task without showing them the way. A best-practice checklist and a user study provide different kinds of evidence.

5. Pick a change you can check

For each observation, write four things: the problem, the evidence, an explanation to investigate and the next check. For example: people leave the form; the drop-off is visible at that step; a required field may be confusing; observe people entering it and try a clearer label.

Do not turn the explanation into a proven result. If traffic or the offer changes alongside the form, a before-and-after comparison cannot isolate the change’s effect. The right measurement approach depends on volume and project constraints.

Start with observations rather than a score

The free UX checklist groups your observations into needs review, not verified, satisfactory and not applicable. It does not scan the site or assign a grade. It makes explicit what you know and what still needs investigating.

If several explanations remain plausible, an audit can help examine the journey and prioritise the investigation. The next step might be a targeted fix, a clearer offer or design work. A full redesign is not the default conclusion.

Sources and further reading