How long should a landing page be: count doubts

A stranger arrives with one or two doubts. Is this the product the ad showed? Is it for someone like me? What does it take to start, and what would I give up to try it? If the first screen answers those, a short page works. If the answer to 'is this for me' only appears after a long section on features, the visitor leaves before they reach it. Length is not the setting to tune first. Put the doubt the visitor brings at the top, answer it in plain words, and then make the page only as long as the next doubt needs.

• The first screen names the product and the person it is for, in words the visitor uses.
• The question from the ad or search is answered before any feature list begins.
• Each section answers one doubt, and its heading names that doubt.
• A claim repeated in a second section adds scrolling and answers nothing new.
• The button repeats what the visitor gets, and it sits after the last answer as well as at the top.

Each traffic source brings a different doubt, and the search phrase shows it. A search that reads 'is X worth it' brings a trust doubt, so the page needs proof and a clear cost before any feature. A search that reads 'X for dentists' brings a fit doubt, so the first screen names the profession. An ad about a limited offer brings a timing doubt, so the offer's end and the next step must show. The wrong version copies the length of a page that ranks for a different question. The right version writes down the doubts for your strongest source, then checks that each one has an answer on the first screen or in the next section.

The longer failure is repetition. A page with three sections that each say the product is easy to use, each with its own button, makes the visitor read the same claim three times without a new answer. The visitor who wanted a new fact gets the same one and stops. The fix is one answer per doubt, placed where the doubt comes up, and one repeat of the button after the last answer. In a simulated run, a visitor who scrolls past the same claim twice and then stops shows the cause: the page repeats itself. A visitor who stops on the first screen shows a different cause: an unanswered doubt.

The report gives each visitor's step and a screenshot, so the stopping point tells you what to change. A visitor who leaves before any scroll never saw an answer to the question from the search or ad, so rewrite the headline and the line under it, not the body. A visitor who stops in a feature section met a section that answers a question nobody brought, so move it below the answer or cut it. A visitor who stops at a form or a button was asked for a decision before the doubt was answered, so put the answer above the form. Change one of those things, then send the same visitors from the same search phrase again.

A team adds a feature tour, a logo wall and a comparison table, and each one looks harmless alone. The answer to the ad's question then drops below the first screen, where a phone visitor may never scroll. Run the same simulated visitors from the same search phrase after each addition, and see whether they still reach the answer. With paid credits, visitors can also reach a page behind a login, using a test account you give them.