A landing page with plenty of clicks and few conversions needs a diagnosis before a redesign. Check whether people actually arrive, whether the offer matches their reason for clicking, whether the page answers their buying questions, and whether the next action works. A weak rate alone cannot tell you which of those transitions failed.
Begin with one commercial action: an order, a qualified enquiry, a booking, or a completed quote request. Button clicks are useful intermediate signals, but they are not substitutes for the outcome. This method concerns acquisition pages and commercial forms; the steps inside a product signup require their own analysis.
Write a conversion contract
Describe the intended journey in a sentence: “A facilities manager looking for a fixed-price maintenance visit lands on this page, checks coverage and availability, and requests a quote.” Include the audience, the immediate need, and the next action. If the team cannot agree on that sentence, settle the disagreement before evaluating the headline.
Define what counts as success and when it is recorded. For an enquiry, distinguish a valid submission from a duplicate, a spam message, and a sales-qualified opportunity. Choose a consistent observation window. A visitor who asks for a quote tomorrow should not automatically be treated as a failed visit today if the normal buying cycle spans several sessions.
Put the denominator beside every rate. “Thirty enquiries from 1,000 eligible page visitors” is much clearer than “3% conversion.” Note exclusions such as employees, automated traffic, and markets the business does not serve. Those are measurement decisions, not cosmetic reporting choices.
Find the first broken transition
Use the sequence below as a diagnostic map. Move from left to right, stopping when you find a reproducible issue. Do not assume every visitor who leaves encountered a fault.
| Transition | Question | Useful evidence | First response |
|---|---|---|---|
| Ad click → page arrival | Did the destination load for this audience? | Landing requests, redirects, device checks | Repair a broken or slow arrival |
| Arrival → offer understanding | Can the visitor identify the offer and who it serves? | Neutral comprehension task | Clarify audience, scope, and outcome |
| Understanding → consideration | Can the visitor resolve the main buying uncertainty? | Questions, comparison behavior, interviews | Supply the missing answer |
| Consideration → action | Does the proposed next step fit their readiness? | Action clicks and task observation | Explain the commitment |
| Action → completion | Can the visitor finish successfully? | Errors, successful receipts, usability tasks | Fix the failed handoff or form |
Check campaign settings alongside the page. A campaign optimized for cheap clicks can bring an audience with a different intent from buyers seeking a quote. This is a hypothesis to investigate through targeting, search terms where available, and actual enquiry quality; it is not a conclusion you can draw from a high click count.
Check arrival on the devices people use
Open the exact campaign URL, including its parameters, on a narrow screen and under a realistic network condition. Check redirects, cookie overlays, sticky elements, consent flows, and the transition to any separate booking or checkout host. Confirm that a successful action remains measurable across the handoff.
For performance, use field measurements where available and reproduce specific problems locally. Google defines Core Web Vitals around loading, responsiveness, and visual stability; these measures help identify technical friction, but they do not explain offer relevance. Use Google’s Web Vitals guide to interpret the metrics rather than treating a performance score as a conversion forecast.
Watch for a layout shift that moves a button just as someone taps it, a hero image that postpones the useful content, or a consent panel that obscures the form. Record the device, browser, page state, and reproduction steps. “Mobile feels bad” is too vague to become a fix.
Compare the promise with the page
Capture the actual advertisement, referring page, or search snippet beside the landing page. Underline what it promises: price, speed, eligibility, delivery, scope, or a particular result. Then identify where the destination qualifies or fulfills each promise.
A promise can fail through omission as well as contradiction. “Book an assessment” leading to a generic “Contact us” form may leave the visitor unsure whether an appointment is reserved or someone will call later. Explain the next step near the action, including any real cost or commitment.
Ask likely buyers to describe the offer in their own words before prompting them. Then give a realistic task: decide whether this service covers their situation and take the next step if it does. GOV.UK’s usability-testing guidance recommends tasks that set a goal without revealing how to complete it. A question such as “Can you see our clear pricing?” teaches the answer and defeats the exercise.
Inspect the decision and the handoff
Make a list of unresolved buying questions, not a list of missing page sections. A visitor might need to know whether an installer works in their building type, whether delivery arrives before an event, or whether a quote requires a paid inspection. The answer could belong in the headline, a short explanation, a diagram, or the form introduction. A testimonial will not resolve a coverage question.
Review the action itself. Use a descriptive label, explain what follows, and ensure the surrounding text accurately describes the commitment. Test completion with a keyboard as well as a pointer. W3C’s forms tutorial provides a useful accessibility reference for labels, instructions, and feedback. Accessibility review is part of checking whether the task can be completed, not an optional visual polish pass.
Worked example: a repair that begins with scope
Synthetic example. A maintenance company receives 1,000 eligible landing-page visits, 120 quote-button clicks, and 30 valid requests. The observed completion rate after a button click is 25%; the overall request rate is 3%. Neither figure explains the cause.
The team reproduces a working form. In task sessions, several participants cannot tell whether apartments are eligible because the page repeatedly mentions offices. Reviewing the campaign shows that apartment-building enquiries were intentionally included. The first proposal is therefore a coverage explanation with concrete eligible building types, followed by a form introduction that describes the quote process.
The team keeps form-error monitoring in place and measures valid requests and enquiry quality after the change. It does not claim that the wording caused improvement without an appropriate comparison. If apartment enquiries prove unsuitable, the remedy may be targeting and qualification rather than more persuasive copy.
Before approving the repair, assign an owner for both implementation and measurement. Record the original page and the exact change so the comparison can be interpreted later.
Worksheet: landing-page diagnosis
Print this page and complete one row for the journey you are reviewing.
| Field | Your answer |
|---|---|
| Audience and immediate need | ______________________________ |
| Exact source promise | ______________________________ |
| Commercial outcome and denominator | ______________________________ |
| First suspicious transition | ______________________________ |
| Observed problem, with device or task | ______________________________ |
| Alternative explanation | ______________________________ |
| Smallest proposed repair | ______________________________ |
| Verification and outcome check | ______________________________ |
What this method cannot establish
A handful of observed sessions can reveal a real usability problem, but cannot estimate its population frequency reliably. Campaign mix, seasonality, inventory, pricing, and sales response time can change outcomes while the page stays identical. Keep those explanations visible. The useful result of diagnosis is a specific, testable account of where the journey fails, with enough evidence to choose the next action.