1. The intent is wrong before the page loads
If a page ranks for a research query and asks for a demo, the conversion rate is telling you the truth. Segment conversion rate by landing page and source before drawing conclusions about the page itself.
2. The message doesn't match the question
Most pages describe the company. Visitors want to know whether this solves their specific problem. The first screen should restate the visitor's problem in their words and name the outcome. The single highest-leverage change on most pages.
3. Proof arrives after the ask
When results, reviews, and specifics sit below the fold while the form sits above it, people are asked to commit before they have a reason to. Move the evidence next to the decision.
4. Friction in the form or checkout
- Fields you don't act on. Every one costs conversions.
- Validation that rejects valid input, especially phone numbers.
- Commitment asked for before any value is delivered.
- No indication of what happens after they submit.
5. The mobile experience is a different product
Most traffic is mobile, and most pages are designed on a large screen. Load the page on a phone on a mediocre connection and try to convert. Sticky calls to action, real tap targets, and lighter pages recover quiet losses.
6. It's converting and you can't see it
Broken or duplicated GA4 events, unmeasured phone calls, form submissions that fire no event, and cross-domain sessions splitting in two can all understate reality. Verify measurement before redesigning anything. The alternative is optimizing against a number that isn't real.
The order to work in
- Confirm tracking is accurate.
- Segment by source and landing page to isolate where the loss is.
- Watch ten session recordings on the worst-performing high-traffic page.
- Fix message and proof first, then form friction, then layout.
- Measure against the previous baseline, and roll winners across templates.