Landing pages that convert: every element from the ad click to the form

Your Google Ads account shows a steady stream of clicks, and the cost per click looks reasonable. Then you count the enquiries at the end of the month and nothing has moved: a few quote requests, two wrong numbers and someone asking about a job. The agency says it is “optimising the campaigns”. You are wondering whether to cut the budget.

In most cases like this the ad isn’t the problem. The page the click lands on is. Below we take a converting landing page apart: message match, the first screen, proof, the form, objections, mobile and speed, trust signals and, finally, testing, the part most teams get wrong. For each one we describe the mistake we see most often and what to do instead.

Message match: send visitors through the right door

The most expensive mistake is also the simplest: sending every ad to the homepage. A homepage is built to tell everyone everything. Someone who searched for “emergency boiler repair Leeds” wants one answer, and if they can’t find it within seconds they hit the back button.

A landing page serves a single goal: a quote request, a booking, a trial, a brochure download. More than one goal means more than one page, which in practice means a page for each ad group or search intent.

Message match means the page answers the ad in the same words. If the ad says “survey within 48 hours”, the headline says 48 hours too. With Meta or YouTube ads the match extends to the visuals: the product in the ad is the product on the page.

Google measures this as well. Keyword-level Quality Score in Google Ads has three components: expected click-through rate, ad relevance and landing page experience, which Google’s help pages describe as how relevant and useful your page is to the people who click. Quality Score isn’t an input to the auction; it is a diagnostic. But if landing page experience reads “below average”, your conversion data almost certainly shows the same problem.

The first screen: four questions in five seconds

What visitors see before scrolling has to answer four questions. What is this? Is it for me? Why you? What do I do now? Leave one unanswered and people either scroll looking for it or leave. Most leave.

The first screen needs only a few parts:

• A headline naming what the visitor wants and how you differ. “Building the future together” is a slogan; “CE-marked warehouse racking, made to your measurements” is a value proposition.
• A subheading adding what the headline couldn’t: who you serve, where, or what is different about your process.
• One primary button, clearly worded and the most prominent thing on the page.
• A short trust line such as an accreditation or a recognisable client. Only what you can prove.
• An image that works as evidence: your premises, your team, your product. Smiling stock-photo office workers convince nobody.

A simple test for value propositions: swap your company name in the headline for a competitor’s. If the sentence is still true, it isn’t a value proposition. “Quality service, competitive prices, satisfied customers” fits every business, which is why it says nothing.

Skip the rotating slider. Auto-advancing slides interrupt reading, split your message three ways and slow the first screen down. One strong message beats three taking turns.

Proof: show what is real, never invent it

Every claim on a landing page raises a silent “really?”. Proof answers it, and the proof that works has one thing in common: it can be checked.

• Genuine reviews from your Google Business Profile or an independent platform, with dates and names.
• Client logos and testimonials used with written permission.
• Short case examples: the problem, the work, the outcome.
• Accreditations (ISO, CE, trade bodies) with validity dates.
• The names, roles and photos of the people who will do the work.

Invented reviews, unsourced “10,000+ happy customers” counters and names attached to stock photos make a page look full for a while and then eat away at trust. A visitor who spots one fake review stops believing the true things too.

It is also a legal matter. In the UK, the Digital Markets, Competition and Consumers Act 2024 has banned fake reviews since 6 April 2025, and the CMA expects anyone publishing reviews to take reasonable steps to keep fakes out. In the EU, the Unfair Commercial Practices Directive, as amended by the Omnibus Directive, blacklists fake reviews and requires traders to say whether and how they check the reviews they show. In the US, the FTC’s rule on reviews and testimonials, in force since October 2024, explicitly covers AI-generated fakes. Using a customer’s name or photo is also personal data processing, so get permission in writing.

Discount claims belong here too: for EU consumers, the “was” price in a price reduction generally has to be the lowest price you charged in the previous 30 days. Check the rules for your market with your legal adviser.

The form and the button: where the enquiry happens

Every field costs the visitor time, attention and sometimes privacy. So ask of each one: could we have the first conversation without it? If yes, drop it or make it optional.

Which fields, and how many?

A service business usually needs a name, a phone number or email, and a short description of the need. Budget range and “how did you hear about us” can wait for the first call.

Long forms aren’t always bad. For an industrial manufacturer, a drawing upload field separates serious enquiries from idle ones. The real question is which enquiries you want. Splitting a long form into steps, the need first and contact details second, usually feels lighter than ten fields at once.

Small technical details matter:

• Use type="tel" and type="email" so mobile users get the right keyboard.
• Set autocomplete attributes so browsers can fill in name, email and phone.
• Put labels above fields, not inside as placeholder text that vanishes once typing starts.
• Show errors next to the field and say what to do: “Please enter a UK mobile number”, not “Invalid input”.
• Send people to a separate thank-you page. It tells them what happens next and gives you a reliable conversion signal for Google Ads and GA4.

The button and the second path

Button text should state the outcome. “Submit” is an action; “Book my site survey” is a result. Keep one primary action and repeat it on long pages. For visitors who aren’t ready, offer a quieter second path: a phone number, WhatsApp or a brochure.

Privacy notice and marketing consent

A form collects personal data, so your transparency duties under the GDPR or UK GDPR start here: put a working privacy notice link next to it. If you only use the data to answer the enquiry, you usually don’t need a consent box for that, and asking for consent you don’t need just confuses people.

Marketing is different. Under PECR in the UK and the ePrivacy rules in the EU, marketing emails to individuals generally need prior consent, the UK “soft opt-in” for existing customers being the main exception. Use a separate, unticked box and never make the enquiry conditional on it. Draft the wording with your legal adviser; our job is to make it work properly on the form.

Answer the objections on the page

Sales calls raise the same questions every time. How long does it take? What affects the price? How does working with you go? What if something goes wrong? If the page is silent, visitors either phone (the good outcome) or go to the competitor whose page answers.

Ask your sales team for the five objections they heard most in the last three months. Most of what follows the first screen comes from that list:

• A “how it works” section in three to five steps.
• What drives the price. Even without published figures, explain what a quote depends on; total silence on price reads as “we’re expensive”.
• A short FAQ of five to eight real questions.
• Risk reducers: a trial lesson, a site visit, clear returns and cancellation terms. If you offer refunds, state the conditions.
• An honest comparison of approaches, off-the-shelf against bespoke for example, without naming competitors.

The table brings together the mistakes we see most when reviewing landing pages, and what we recommend instead.

Mobile, speed and trust signals

A large share of ad traffic arrives on phones; your ad account’s device report gives you your own figure. Design for the phone first, then expand for desktop.

• Keep the primary button within thumb reach. A button fixed to the bottom of the screen works well on long pages, as long as it doesn’t cover content.
• WCAG 2.2’s AA-level target size criterion asks for at least 24 by 24 CSS pixels; for primary buttons, the 44 to 48 pixels in Apple’s and Google’s interface guidelines are more comfortable.
• Phone numbers should dial when tapped and addresses should open in a map app.
• Check that the cookie banner and chat bubble don’t cover the button or form, on your own phone, on mobile data, as a first-time visitor.

Speed thresholds are covered in our Core Web Vitals article. In short: the large first-screen image should load quickly (Google treats a Largest Contentful Paint of 2.5 seconds or less as good), the layout shouldn’t jump, and the button should respond instantly. What slows landing pages most is rarely the design but years of piled-up tracking code: old pixels, tags installed twice, chat tools nobody uses.

Trust signals sit at the bottom where everyone forgets them, yet visitors check them right before filling in the form:

• Your registered company name, address and phone number.
• A working, current privacy notice link next to the form.
• Cookie preferences that can be changed at any time.
• A valid HTTPS certificate. A “not secure” warning on a form page kills the enquiry.
• For online shops, delivery, returns and cancellation terms one click away.

Testing: doing A/B tests properly

An A/B test splits traffic randomly between two versions over the same period and measures which converts better. It sounds simple, which is exactly why it gets botched so often.

First question: do you have enough traffic? With a few dozen conversions a month, separating small differences with any confidence takes months. Customer interviews, session recordings, a five-person usability test and bold changes will get you further, faster.

If you do test, follow these rules:

1. Start with a hypothesis: “Putting the timescale in the headline will increase form completions, because timing is the question sales hears most.”
2. Calculate the sample size first. Enter your conversion rate, the smallest difference you care about and your confidence level into a sample size calculator.
3. Fix the duration in advance and run full weeks. A Monday visitor isn’t a Saturday visitor; our minimum is two full weeks.
4. Don’t stop early. Checking daily and stopping the moment something looks significant is the quickest way to pick the wrong winner.
5. Check the split. If a 50/50 test runs 58/42, the set-up is broken and the result can’t be trusted.
6. Don’t change budgets, campaign structure or pricing mid-test.
7. Roll out the winner and write it down. Losing tests are information too; teams without notes rerun them a year later.

Google Optimize closed in September 2023. Teams now use tools such as VWO, Optimizely or AB Tasty, or split traffic on the server. The tool matters far less than the rules.

Heatmaps and session recordings: consent first

Tools such as Microsoft Clarity or Hotjar show where visitors click, where they get stuck and which form field makes them give up. On pages without enough traffic for A/B tests, they are often your most useful source of insight.

They also store information on the visitor’s device and process personal data. In the EU, the ePrivacy rules require consent for anything not strictly necessary, analytics included. In the UK, the Data (Use and Access) Act 2025 added a narrow PECR exception for purely statistical analytics, provided people are told and can object easily; the ICO’s storage and access technologies guidance, finalised in April 2026, sets out the conditions. In our view, session replay of individual behaviour fits that “sole purpose” test poorly, so treat it as consent-based unless your legal adviser says otherwise. Clarity itself has required a valid consent signal for visits from the EEA, UK and Switzerland since 31 October 2025.

Examples by sector

The principles stay the same; the execution changes by sector.

Dental practice: implant consultations

A dental group with three practices advertises implants. A good landing page asks which practice first, because visitors want the nearest one. It names the dentists, explains the treatment step by step and keeps the booking form short. Healthcare advertising is tightly regulated: in the UK the CAP Code and your professional regulator’s guidance apply. Review before-and-after photos, testimonials and claims like “pain-free” with your legal adviser first.

Industrial manufacturer: quote requests

Visitors to an exporting machined-parts manufacturer are usually buyers comparing several suppliers. The page needs accreditations, capacity, tolerances and materials, delivery terms (Incoterms if you ship abroad) and a drawing upload. A version for another market should be built for how buyers there purchase, not just translated.

Language school: placement tests

For a language school the strongest conversion point is often a placement test or trial lesson, not direct enrolment. Ask for a small step first. Show course dates, class sizes, teacher profiles and the certificates students can work towards.

Where to start

1. List your five highest-spending ad groups and the page each one sends traffic to.
2. Put each ad’s promise next to the page headline; fix mismatches or build separate pages.
3. Check the first screen against the four questions: what is this, is it for me, why you, what now?
4. Check the source and permission behind every piece of proof; remove what you can’t back up.
5. Run every form field through “could we have the first conversation without this?” and check the privacy link and marketing consent box.
6. Get the five most common objections from sales and answer each on the page.
7. Test the page on your own phone, on mobile data. Submit the form for real and confirm the conversion is recorded.
8. With enough traffic, plan your first A/B test with a hypothesis, a calculated sample and a fixed duration. Without it, start with consent-based session recordings and customer interviews.

At AYTE Digital we don’t treat landing pages separately from the ad account: we look at the data from your current ads and pages, then decide with you which change comes first. If you’d like to go through your page together, a screen-share on the first call is a good place to start.