Why your website is not converting

Diagnose the leak before you redesign

The most expensive mistake in web design is redesigning a site that did not need redesigning. A full rebuild feels like progress, but if the problem was one section — the headline, the pricing, the signup form — a rebuild wastes months and often reintroduces the same problem in a prettier form. Diagnosis comes first, and it is usually fast.

Start with the journey, not the design. Where do visitors arrive? What do they see first? What do they do next, and where do they stop? If most visitors arrive on the homepage and leave without scrolling, the first screen is the leak. If they scroll, read and leave at the pricing page, the pricing story is the leak. If they start a signup and abandon it, the form or the ask is the leak. Each leak has a different fix, and the fix for one will not help another.

The first screen test

Show the first screen to someone who has never seen the site, for five seconds, then hide it and ask three questions: what does this product do, who is it for, and what should I do next? If they cannot answer all three, the first screen has failed, however beautiful it is. This test is brutal and it is almost always right, because it reproduces exactly what a real visitor does — except a real visitor does not get asked, they just leave.

Most first screens fail on the first question. They say things like “empowering teams through innovative solutions” or “the future of work, today” — sentences that could describe a thousand products and therefore describe none of them. A first screen that works says something specific: who it is for, what it does, and what changes for them. Specificity is what makes a visitor stay. Vagueness is what makes them bounce, because a vague headline asks the visitor to do the work of figuring out whether the product is for them — and visitors do not do unpaid work.

Copy: who, what, next — in that order

Good website copy follows a simple order, and most failing sites break it. First, who: the visitor must recognise themselves in the first screen, or they assume the product is for someone else. Second, what: the product, in plain language, without jargon or internal names. Third, next: the one action, stated plainly — start a trial, book a demo, buy, enquire — not “learn more” or “discover”, which are not actions at all.

After that order comes proof, and proof has its own order too. The strongest proof is specific: a named customer, a concrete outcome, a recognisable logo, a number that means something in context. The weakest proof is generic: “trusted by teams everywhere”, “loved by thousands”, “best in class”. Generic proof is worse than no proof, because it signals that there was nothing specific to say. If you do not have specific proof yet, say what you do have — early customers, a pilot, a clear product — rather than inventing a claim. A site that is honest about being new converts better than a site that pretends to be established, because visitors can tell the difference.

The action problem

Many sites do not have a conversion problem at all — they have an action problem. The action is hidden in a menu, stated vaguely, asked too early, or split across five competing buttons. A visitor who wants to act should never have to hunt for how. The action should be visible in the first screen, repeated after the proof, and stated in the same words every time, so the visitor recognises it when they are ready.

Asking too early is the subtler version of the same problem. A site that asks for a credit card, a demo booking and a phone number on the first visit is asking for a commitment the visitor has not yet made. The right ask depends on the product: a trial for a SaaS product, a sample for a physical product, an enquiry for a service. The principle is the same — ask for the smallest commitment that moves the visitor forward, and earn the next one after that. A site that asks for everything at once usually gets nothing.

A hypothetical example

Take a hypothetical SaaS company whose site got steady traffic and almost no trials. The homepage headline was “Work smarter, together” — a sentence that could describe any productivity product — and the first action was “Book a demo”, buried below three sections of feature descriptions. Visitors arrived, did not recognise themselves, did not understand the product, and were asked for a demo they had no reason to want.

The fix, in order, was: a headline that said who the product was for and what it did; a first screen that ended with a trial signup, not a demo booking; proof — a named customer and a concrete outcome — placed before the feature list, not after it; and a demo offered as a second action, for visitors who wanted help, rather than as the only action. No redesign of the whole site was needed. The leak was the first screen, the copy order and the action — and fixing those three, in that order, is what changed the number. That is the pattern for most “website not converting” problems: the leak is specific, and the fix is specific, once you have found it.

What should you change first?

Change in this order, and measure after each change: first the first screen, then the copy order on the pages where traffic actually lands, then the action, then the proof. Only consider a full redesign after those four, and only if the brand, product or audience has changed.

Change in this order, and measure after each change, so you know what worked. First, the first screen: headline, subheadline and action, tested with the five-second test above. Second, the copy order on the pages where traffic actually lands — usually the homepage and one or two landing pages, not the whole site. Third, the action: one action per page, visible, specific, and asked at the right moment. Fourth, the proof: specific, placed before the features, and honest about what exists.

Only after those four should you consider a full redesign — and only if the brand, the product or the audience has changed, not because one page underperforms. If you are not sure which leak you have, the diagnostic table at the top of this post is a good place to start, and the web design service page describes how a redesign is scoped when one is actually needed.