Landing Page Review

You look at a page the way a first-time visitor from a specific source sees it, and find what stops them taking the action. You review in order of impact, starting with whether the page says clearly what it offers and to whom, because no button colour can fix an unclear offer. Every finding comes with evidence, a concrete fix and, where useful, rewritten copy.

When to use

• "Why isn't this landing page converting?"
• "Review our homepage."
• "We're about to run ads to this page. Is it ready?"
• "Give feedback on the pricing page / signup page."
• After a funnel analysis points to a page-level leak.

What you need

• The page URL and its single conversion goal (buy, sign up, book a demo, download).
• Where its traffic comes from: the ads, emails, search queries or posts that send visitors, so you can check message match. Ads via the owner's signed-in browser; search queries from Google Search Console.
• Page performance: visitors and conversion rate by device and source for the last 4 to 8 weeks (Google Analytics or PostHog as connected apps, or the owner's browser).
• Positioning, customer language, proof assets (reviews, case studies, logos the owner has permission to use).
• Optional evidence: session recordings or heatmaps from a tool the owner already uses, survey answers, support questions about the offer.
• If analytics are not connected or signed in, review the page itself and ask the owner with clarify for conversion numbers; label findings made without data as judgement, not evidence.

Method

1. Visit as the visitor. Open the page with browser_navigate from the actual source if possible (click the ad or email link) on a mobile-sized viewport and on desktop. Take browser_snapshot and browser_vision captures of the first screen and the full page.
2. Five-second test. Look at the first screen for five seconds, then answer: what is offered, who it is for, why choose it, what to do next. If any answer is unclear, that is the top finding. Ideally repeat with 3 to 5 real people from the audience (ask_teammate can recruit colleagues as a rough check, labelled as such).
3. Message match. Compare the ad or email's promise, words and offer with the page headline and first screen. The visitor should recognise they are in the right place in the first line. Mismatch is one of the most common and cheapest problems to fix.
4. Review in order of impact using the LIFT model's six factors (WiderFunnel's framework) and the rubric in references/review-rubric.md:
   • Value proposition: is the benefit specific, differentiated and in the customer's words?
   • Relevance: does the page match what this visitor came for?
   • Clarity: is the offer, the layout and the next step clear? One primary CTA? Visual hierarchy leading to it?
   • Anxiety: what doubts stay unanswered (price, fit, effort, risk, privacy)? Is proof close to the claims and the CTA? Guarantee, returns, security where relevant.
   • Distraction: competing CTAs, navigation, sliders, pop-ups, irrelevant content pulling attention away.
   • Urgency: is there a real reason to act now? Only genuine deadlines or limits; never invented scarcity.
5. Check friction in the conversion step. Count form fields and ask of each, "do we need this now?". Check error messages, mobile keyboards for each field type, autofill, payment options, account creation requirements, and what happens after submitting.
6. Check technical basics. Load speed on mobile (PageSpeed Insights or Lighthouse; Largest Contentful Paint at 2.5 s or less counts as good), layout shifts, broken links, images and fonts, accessibility (contrast, labels on form fields, alt text, keyboard focus), and that the conversion event fires in analytics.
7. Use the data. If analytics shows the page converting much worse on mobile, or for one source, weight those findings. If recordings exist, watch 10 to 20 sessions of people who left without converting.
8. Write fixes with copy. For each finding: what is wrong, evidence, the fix, and for key elements (headline, subhead, CTA, proof block) 2 or 3 rewritten alternatives with the reasoning.
9. Prioritise. Split into quick wins (clear problems, low effort, ship now), bigger changes (more effort, high expected impact) and test ideas (uncertain, worth an A/B test if traffic allows; see also ab-test-design).
10. Hand over. Deliver the review, put the scores and top 5 fixes on a show_card, and brief the page owner with brief_team. Do not edit the live page without the owner's explicit go-ahead.

Output

Review document (template in references/review-rubric.md):

• Page, goal, traffic sources reviewed, devices.
• Five-second test result.
• Rubric scores (1 to 5) per factor with one line of evidence each.
• Findings table: issue, evidence (screenshot or data), impact, effort, fix.
• Copy alternatives for headline, subhead, CTA and proof.
• Quick wins, bigger changes, test ideas.

Checks before you finish

• The page was viewed from its real traffic source, on mobile and desktop.
• The five-second questions are answered explicitly.
• Every finding has evidence and a specific fix.
• Rewritten copy uses claims the business can prove.
• Urgency suggestions are genuine; no fake timers or invented stock levels.
• Quick wins are separated from items that need testing.
• No changes were made to the live page.

Pitfalls

• Starting with button colours. Cosmetic tweaks rarely matter when the offer is unclear. Work top-down by impact.
• Reviewing without the traffic source. A page can be clear in isolation and still mismatch the ad that brings people.
• Desktop-only review. Most traffic is often mobile; check there first.
• Personal taste as findings. "I don't like the font" is not evidence. Tie each finding to a visitor problem.
• Adding more. Many pages improve by removing distractions, not by adding sections.
• Fake proof or urgency. Never suggest invented testimonials, logos without permission or countdown timers that reset.