Inside LAND: ranking landing pages by wasted traffic

Short answer: some landing pages earn the traffic you send them. Others quietly burn it. LAND takes your GA4 landing page numbers, ranks every row by a waste score so the worst sits on top, then live-checks the worst 20 to tell you whether the problem is technical, content or a promise the page doesn't keep.

The usage guide is LAND: how it works. This is the scoring.

What you paste in

Rows of URL, sessions, engagement rate (or bounce rate) and conversions, from GA4's Landing page report. CSV, TSV or pipe-separated rows all work, with or without a header. With a header, LAND finds columns by name, and "key event" counts as conversions. Without one, the order is URL, sessions, engagement, conversions. You can paste up to 50 unique URLs, and only full http or https URLs count.

The waste score, under the hood

Every row gets one number:

• +1000 if the page had sessions and zero conversions. That outranks everything else on purpose. Traffic with no outcome is the clearest kind of waste.
• +2 for every point of engagement below 100%. If you only gave bounce rate, engagement is taken as 100 minus bounce. With neither, it's 50.
• + sessions / 1000, capped at 10. Volume breaks ties, but it can't outweigh the other two.

Rows are sorted by score, and the top 20 are fetched live. The rest are listed as "not live-checked this run", so you can re-run on a smaller list.

The live verdict, in order

For each fetched lander, the first match wins:

1. Broken (P0, Eng): 5xx, 404, 410 or 403. Traffic is landing on a broken page, so fix that before any copy work.
2. noindex (P1, Eng): fine for paid-only landers, wrong if you expect organic traffic.
3. Redirects (P1, Eng): point ads and links at the final URL so the numbers line up.
4. No visible next action (P0, Growth): no form, button, CTA-classed link or submit input in the source.
5. Thin (P1, Content): under 150 words.
6. H1 count isn't one (P1, Content).
7. Sound but zero conversions (P0, Growth): check the offer, the match to the ad or query, and that the conversion event actually fires.
8. Engagement under 40% (P1, Growth): a match problem. Check what people expected before they landed.
9. Otherwise (P2): leave it alone.

Edge cases

• "No CTA" means none in the server HTML. A button injected later by a script won't be seen. If the page clearly has one in the browser, it's probably rendered client-side. Treat that as a ticket for Eng, not for Growth.
• Word count includes navigation. A page can pass 150 words on menu text alone. A pass means "not obviously empty".
• Zero conversions might be zero tracking. Before you rewrite a page that "converts nobody", run VERA on it to prove the conversion event is even on the page.

Worked example

url,sessions,engagement_rate,conversions
https://samplebrand.test/lp/spring-kits,1200,31,0
https://samplebrand.test/kits/blue-kit,4800,62,22
https://samplebrand.test/lp/trade,650,28,3

Waste scores: 1139.2 | 80.8 | 144.65
Live: /lp/spring-kits 200, 1 x h1, 420 words, form present

The spring lander ranks first by a mile: 1,200 sessions and no conversions. It's technically sound, so the problem is the offer, the ad match or the tracking. Run VERA before anyone touches the copy. /lp/trade comes second on weak engagement. The blue kit page, which is doing its job, sits at the bottom where it belongs.

Run it: open LAND

The money angle

Waste is the easiest revenue to recover, because you've already paid for the visit. LAND puts the pages where that spend leaks at the top of the list, and the order of its verdicts stops the classic mistake: rewriting the copy on a page whose form doesn't submit.

When a traffic drop turns out to be a conversion drop, this is where the traffic drop audit goes next. Value the top rows in PAYBACK.