Landing page fixes that lift conversion

Most landing pages don’t lose sales because they look dated. They lose sales because a visitor on a phone can’t tell, within a few seconds, whether this is the thing from the ad, what it costs and why they should believe it.

Here are the fixes we check first, roughly in order of how often they matter and how cheap they are to ship.

The fix list, in priority order

1. Match the message to the ad. If the ad says “first box €29,” the page should say “first box €29” above the fold, with the same product and the same visual. A visitor who clicked on one promise and lands on a generic homepage has to start over. Build one page per major ad angle, not one page per ad.

2. Make the offer clear above the fold. On a phone, without scrolling: what it is, who it’s for, the price or starting price, and one call to action. Show the page to someone for five seconds and ask what it sells. If they can’t answer, fix that first.

3. Speed. Test on a mid-range phone on a mobile connection, not on office Wi-Fi. Compress and resize images, lazy-load everything below the fold and remove scripts from apps you no longer use. A useful target: headline and main image visible in about 2.5 seconds or less.

4. Put social proof where the doubt is. A star rating next to the price. Two or three reviews near the add-to-cart button that answer the usual objections: texture, sizing, how long results take. A logo wall at the bottom does little for someone deciding at the top.

5. Cut form and checkout friction. Offer guest checkout and express wallets. Remove fields you don’t need. Don’t force account creation before payment. For B2B lead forms, ask only for what sales needs to qualify. Small rate changes compound: in our Kelpie Skin demo case (illustrative figures; results vary), add-to-cart to purchase went from 32% to 38%, which on about 11,700 monthly carts is roughly 700 more orders from the same carts.

6. Design for the thumb. Sticky add-to-cart on mobile, tap targets large enough to hit, readable body text, details in collapsible sections, and no popup in the first few seconds. Most paid social traffic arrives on a phone, so review the page on one.

7. Be clear about price and shipping. Put the shipping cost or free-shipping threshold and the delivery time next to the price, and the returns policy near the button. For cross-border orders, say who pays duties and taxes. Costs that first appear at checkout send people away.

Run tests with enough traffic

A test that ends too early is worse than no test, because it gives you false confidence. A rough rule for visitors needed per variant, at 80% power and 5% significance:

Visitors per variant ≈ 16 × p × (1 − p) ÷ d²

Here p is your current conversion rate and d is the absolute change you want to detect. Say you convert at 3% and want to detect a lift to 3.6%, a 20% relative improvement. That’s 16 × 0.03 × 0.97 ÷ 0.006², or about 13,000 visitors per variant. At 1,000 visitors a day split two ways, that’s close to four weeks.

To detect a 10% lift, you need about four times as many. That’s why low-traffic pages should test big changes.

Rules we don’t bend:

• Set the sample size before you start, and don’t stop the moment a result looks significant.

• Run in full weeks, at least two, so weekdays and weekends both count.

• Pick one primary metric, conversion rate or revenue per visitor, before launch.

• Don’t change ads, offers or prices for the tested traffic mid-test.

Score the backlog

You’ll have more ideas than traffic. Score them so the order isn’t set by whoever talks loudest.

ICE is the simplest: rate Impact, Confidence and Ease from 1 to 10 and average them. PXL-style scoring swaps gut ratings for yes/no questions. Is the change above the fold? Is it noticeable within five seconds? Is it based on something you’ve actually seen in recordings, surveys or analytics? Yes/no answers are harder to inflate.

The redesign and the button color tie at the bottom, for opposite reasons.

What not to test first

• Button colors and small copy tweaks. Cheap to run, rarely big enough to detect on normal traffic.

• Full redesigns. Too many changes at once. Even if one wins, you don’t know why.

• Low-traffic pages. Start where the visitors are, usually your top two or three paid landing pages.

• Bugs. A broken mobile layout or a six-second load isn’t a test. Fix it and move on.

• Price tests before clarity fixes. If visitors don’t understand the offer, a price test measures confusion.

If you’d like a ranked list for your own pages, our Growth Audit ends with a 90-day plan that includes one.