Five landing-page tests worth running this quarter.

Most landing-page tests we inherit are button colours and hero photos. They're easy to set up and almost never move anything that matters. The tests worth your traffic change what a visitor understands, trusts or has to do.

These are the five we'd run first on almost any lead-generation or campaign page, in roughly this order. Each is a single, clear change, so you'll know what caused the result.

One rule before any of them: testing is for questions you can't answer by looking. A broken form, missing conversion tracking, a page that takes seconds to show anything on mobile, a headline that contradicts the ad — fix those first. A test on a broken page only tells you which broken version is less bad.

Visitors arrive with the words of an ad or a search in their heads. If the headline says something different, they have to work out whether they're in the right place. Test a variant whose headline repeats the promise of the ad or keyword group sending the traffic, in the same language. For campaigns with several ad groups, that may mean one page with a few headline variants rather than one headline for everyone. Keep the variant honest: if the ad promises a price, a timeframe or an offer, the page has to show it on the first screen, not three scrolls down.

Every field is a small reason to leave. List each field on your form and ask the sales team which ones they actually use to qualify a lead, then test a version with only those. If you genuinely need more, test a two-step form: easy questions first, contact details second. The goal is qualified leads, so measure what happens after the form, not just how many people submit it. Phone number is the field we see cut most often; if sales calls everyone back, keep it and say why you're asking.

Testimonials and client logos often sit in a band halfway down the page, far from the moment of decision. Move one specific piece of proof — a short quote with a name and role, your real review rating and count, a relevant result — directly beside the main call to action. Specific beats generic: a named customer and a concrete outcome do more than “Trusted by thousands”. In our experience business buyers lean on recognisable client names and consumers on review ratings, so test which kind of proof your audience trusts.

“Submit” and “Learn more” tell visitors nothing. Test call-to-action copy that names the outcome and removes uncertainty: “Get my quote”, “Book a 20-minute call”, “See pricing”. Pair it with a short line underneath that answers the obvious worry — no card required, a reply within one working day, no obligation — as long as it's true. Then make sure the next screen delivers exactly what the button promised.

Hero videos, sliders and large background images often push the headline and button down the page and slow the first render, especially on mobile. Test a variant with a static, compressed image (or none) and the offer and button visible on the first screen of a phone. If your paid traffic is mostly mobile, this is often the test we'd run first, and our 1 MB page budget explains how to keep the page light afterwards.

• Choose the primary metric before you start — ideally qualified leads or sales, not clicks.
• Estimate the sample size you need up front, and don't stop the test the moment one version pulls ahead.
• Run for whole weeks, so weekday and weekend behaviour are both included.
• Change one thing per test. Change everything and you'll learn that something worked, but not what.
• Write down the result and why you think it happened, win or lose. The notes outlast any single test.

When a variant wins, roll it out to all traffic and check the numbers again a month later. Wins often shrink after launch. That's normal, and it's why we keep a log of results and reasons rather than a list of trophies.

Any of these can be set up in a dedicated testing tool or with server-side feature flags. What matters more is the discipline: one change, one metric, enough time. Our digital marketing team runs this programme alongside paid campaigns, so tests and traffic are planned together — and the attribution clean-up comes first, so the results can be trusted.

Written by

Lena Mårtensson

Partner · Design · 13 years in product design

Lena leads design across our websites, apps and campaigns. She cares about what a visitor understands in the first five seconds, and writes here about design decisions you can test rather than argue about.

LinkedIn