Why your website isn’t converting — seven causes, in fixing order.

A site with visitors and no enquiries is a specific, diagnosable problem, and the instinct to redesign it is usually wrong — a redesign changes everything at once, which means you learn nothing and may break something that worked. These are the seven causes we find most often in conversion work, in the order that returns the most for the least effort.

Before anything: find out where people leave

Without data you are redecorating. You need three things: how many people arrive, how many reach the point of enquiry, and how many complete it. GA4 with properly configured conversion events answers the first and third; a scroll or funnel view answers the second. It is an afternoon of setup and it turns the rest of this list from opinion into a priority order.

The most common discovery is that the page is not the problem at all — the traffic is. Visitors arriving on a keyword that does not match what you sell will never convert, no matter what the button says. Check intent before you touch layout.

Cause 1: the page does not say what you do

The single most common failure. A visitor lands and cannot answer, within a few seconds, what you sell, who it is for and why you rather than the other tab they have open. Headlines built from ambition — "empowering businesses to thrive digitally" — read as noise, because they could belong to any company on earth.

The fix is unglamorous: a plain sentence naming the thing and the buyer. "Fixed-price websites for small businesses in Bucharest and Berlin" converts better than any adjective, because it lets the wrong visitor leave and the right one recognise themselves.

You have less time to do it than the layout implies. Nielsen Norman Group’s analysis of page-visit durations found users typically leave within 10–20 seconds, and that pages holding attention for minutes are the ones that communicated a clear value proposition inside that first window. The average visit is under a minute. Whatever a visitor has to scroll to understand, most of them will not.

Cause 2: no price, and no reason for that

Hiding pricing is a decision to lose every visitor who is not yet ready to talk to a human — which, on a first visit, is most of them. Buyers are trying to establish whether you are in their range at all, and "contact us for a quote" reads as "expensive and evasive" to exactly the people you can afford to serve.

If real prices are impossible, publish the shape of them: starting points, typical ranges, what drives a number up. Our prices are fully public for this reason, and the price guide exists to show the working. Anything is better than a blank.

Cause 3: one call to action too many

A page offering "book a call", "download the guide", "see our work", "join the newsletter" and "get a quote" with equal visual weight is asking the visitor to make a decision they have no basis for making. Competing actions do not add up; they cancel out.

One primary action per page, repeated as the visitor scrolls, with everything else visibly secondary. If two actions genuinely matter, make one clearly dominant — the person who wants the other will still find it.

Cause 4: the form asks for too much

Every field is a chance to leave. Company size, job title, phone number, budget dropdown, "how did you hear about us" — each one is a small tax on someone who has already decided to contact you, and phone number is the most expensive field on the internet.

Ask for what you need to reply: an email address and a message. Name should never be required. Qualification questions belong in your reply, not in the gate — and if you must ask, use optional pills rather than mandatory dropdowns.

Causes 5 and 6: it is slow, and it is awkward on a phone

Speed is a conversion feature, not a technical vanity metric. A hero image that arrives two seconds late, a layout that shifts as fonts load, a button that does nothing for half a second after being tapped — each loses a measurable share of visitors before they see your offer. Core Web Vitals exist precisely because these three things predict abandonment.

Mobile compounds all of it. Most traffic is a phone, and the failures are mundane: tap targets too close together, a sticky header eating a third of the screen, a form that scrolls the keyboard over the field being typed into, horizontal overflow from one wide table. Test on a real device, not a resized browser window.

Cause 7: nothing on the page makes you credible

Strangers are being asked to send money or personal details to a company they have never heard of. Anonymous stock photography, no address, no named people, no real work shown, testimonials attributed to "J.D., CEO" — all of it reads as risk. Trust is not a section at the bottom of the homepage; it is the accumulated effect of specifics.

Named clients, real project outcomes, a physical address, faces, a phone number that connects to someone. Our portfolio and contact page are built as trust surfaces rather than decoration, because they are what a cautious buyer actually reads.

The order to fix them in

Measurement first, then clarity of message, then pricing visibility, then the form, then speed, then mobile, then trust. That order is roughly cheapest-and-highest-impact to most-effort, and clarity of message alone frequently moves the number more than everything below it combined.

If you would rather have someone else do the diagnosis, our UX audit is free and produces a ranked list rather than a sales pitch. If the conclusion is that the structure itself is the problem, a conversion redesign is the scoped version of that — but we would rather rule out the cheap fixes first.

One calibration before you start: some of what looks like failure is just how the web behaves. Baymard Institute’s meta-analysis of 50 studies puts average cart abandonment at 70.22%, and roughly half of that is people who were never going to buy on that visit. The goal is to move your own number, not to reach a rate nobody achieves — and if you sell products rather than services, checkout abandonment breaks that 70% down into the ten reasons shoppers actually give.

Cause 8: the form fights the phone

A short form can still be a slow one. Cause 4 was asking for too much; this is asking for the right things badly. If the inputs carry no autocomplete attribute, the browser cannot offer details it already holds, so every visitor retypes their name, email and address with a thumb. MDN puts the accessibility case plainly — appropriate autocomplete values help people with cognitive and motor impairments fill forms “more quickly and accurately by reducing the need to type and remember information” — and the same change quietly speeds up everyone with a saved profile.

The fix is unglamorous and takes an afternoon: autocomplete on the right fields (given-name, family-name, email, tel, postal-code), and the right input types so a phone offers a keyboard with an @ on it instead of the full alphabet. Nobody will ever compliment you on this. What you will see is fewer half-finished submissions, because you removed several seconds and a dozen taps from the one moment when the visitor had already decided to say yes and just needed the page to get out of the way.

Sources

The behavioural benchmarks and the form-input guidance above come from these, checked August 2026. The eight causes and the fixing order are what we find in our own audits.

• Nielsen Norman Group — How Long Do Users Stay on Web Pages? ↗

  Analysis of page-visit durations by Jakob Nielsen: users often leave within 10–20 seconds, the average visit runs a little under a minute, and pages that hold attention longer are the ones stating a clear value proposition inside the first 10 seconds.

• Baymard Institute — Cart abandonment rate statistics ↗

  Average documented cart abandonment of 70.22%, calculated across 50 separate studies — the benchmark to judge your own rate against rather than an aspiration to eliminate.

• web.dev — Web Vitals ↗

  The definitions and thresholds behind causes 5 and 6: LCP under 2.5s, INP under 200ms, CLS under 0.1, measured at the 75th percentile of real visits.

• Analytics Help — [GA4] Engagement rate and bounce rate ↗

  What GA4 actually counts as an engaged session — over 10 seconds, or a key event, or at least two page views — which is the measurement the diagnosis step above depends on.

• MDN Web Docs — The autocomplete attribute ↗

  Appropriate autocomplete values let the browser fill known details and help users with cognitive and motor impairments complete forms more quickly and accurately, reducing typing and recall.