Most advice about SaaS landing pages comes from someone's opinion or from a vendor's own customers. We wanted the third thing: a neutral sample of real companies, scored the same way, with the raw data published so you can check us.
On 25 July 2026 we crawled 800 company sites drawn from Y Combinator's public directory, filtered to active B2B and SaaS companies, shuffled with a published seed (20260725) so anyone can rebuild the same sample. 675 returned a page we could score. The other 125 failed in ways we disclose below, because dropping them would have flattered every number here.
The headline finding
30.7% of the 675 pages had no call to action in the body of the page.
That figure excludes the navigation, the header and the footer. A sticky header with a "Get started" button appears on almost every SaaS site, and counting it would measure the existence of a header rather than the work the page does. Strip those out and close to a third of these companies ask the visitor for nothing.
These are funded startups. They have money, designers and a reason to care.
The median SaaS landing page
| What we measured | Median | p25 | p75 | n |
|---|---|---|---|---|
| Word count | 759 | 458 | 1,218 | 675 |
| Words before the first in-body CTA | 34 | 21 | 101 | 464 |
| Reading grade (Flesch-Kincaid) | 10.4 | 8.3 | 12.5 | 595 |
| Distinct CTA labels | 1 | 0 | 2 | 675 |
| Total CTA elements | 2 | 0 | 3 | 675 |
| Form fields | 0 | 0 | 1 | 675 |
| Images with no alt text | 6 | 0 | 34.5 | 675 |
Denominators differ per row. Reading grade only scores genuine prose, so pages that are mostly interface chrome return nothing. The words-before-CTA row only counts pages that had a CTA to measure against.
A note on that row: 34 words is a document-order position, not a claim about the fold. Measuring the fold needs a rendered viewport at a known screen size, which our crawler does not do. Anyone who tells you what percentage of pages put the CTA below the fold, without saying which viewport they used, is guessing.
Four numbers worth acting on
42.2% showed any social proof. A logo wall, a testimonial, a customer count, a review score. Any of it counted. Fewer than half had one.
11.6% put social proof before the first call to action. So even among the pages that carry proof, most ask the visitor to act before giving them a reason to trust the company asking.
That ordering matters more than the presence. Ziakis and colleagues surveyed 184 people on what drives their decisions on a landing page and found trust ranked above every other dimension they tested, with user reviews, guarantees and detailed descriptions all scoring above 78%. Personalization ranked last. Their sample was a convenience sample recruited over social media, so treat it as a signal about stated preference rather than a measurement of behaviour. Even discounted, it points the same way: visitors want a reason to believe you, and 88.4% of these pages ask them to act before supplying one.
15.4% stated a price. The other 571 companies want you to click through, book a call, or work it out yourself.
8.1% shipped no H1 at all, and 7.1% of the 620 pages with a headline used one longer than ten words.
What broke, and why we are telling you
125 of the 800 pages never produced a score:
| Failure | Count |
|---|---|
| Served no meaningful HTML without JavaScript | 83 |
| Network error | 13 |
| HTTP 403 | 8 |
| Body too small to score | 7 |
| HTTP 429 | 5 |
| HTTP 404 | 4 |
| Robots disallowed | 2 |
| Other (503, 526, 402) | 3 |
The first row is the interesting one. 83 companies served a page that was empty until a JavaScript bundle executed. A crawler that does not run scripts sees nothing. Some search crawlers render and some do not, and no AI retrieval system guarantees it.
Most published studies drop these pages. We scored them as failures rather than zeros, because a scored empty page drags every median toward zero and makes the whole dataset useless.
Where our data stops
We measured what these companies ship. We did not measure what converts, and no number on this page should be read as a conversion claim.
For that, Unbounce has the better dataset: roughly 41,000 landing pages, 464 million pageviews and 57 million conversions. Their finding on calls to action is worth putting beside ours. Pages with a single CTA link converted at 13.5%, pages with two to four at 11.9%, and pages with five or more at 10.5%. They also found 14.8% of pages used a single CTA link while 68.2% carried more than five.
Their population differs from ours in a way that matters. Every page in the Unbounce sample was built by someone who bought a conversion optimization tool. That is a selected group, and it generalises to people who buy CRO tools. Our frame is YC startups selected by a random seed, which generalises to startups. Both are useful. Neither substitutes for the other.
How to use this
Open your own page and check five things against the table above.
Does the body of the page ask for anything, ignoring the header? Does proof appear before the ask? Is there a price? Does an H1 exist? Does the page render with JavaScript disabled?
If you clear all five you are ahead of most of this sample, which is a lower bar than it sounds.
Method and data
Sample frame: Y Combinator public company directory via the yc-oss mirror, filtered to active companies classified B2B or SaaS. Seeded shuffle, seed 20260725. 800 attempted, 675 scored, collected 25 July 2026. Robots.txt honoured, identifying user agent, bounded concurrency.
Two metrics named in our original plan are not implemented and appear nowhere above: whether the primary CTA sits below the fold, and whether the page breaks at 390px. Both need a rendered viewport. We say so in the output file itself.
The full dataset, the per-metric denominators and the failure ledger are published with the study.
Sources
- Ziakis, C., Papadopoulos, A., Antoniadis, I. and Saprikis, V. (2026). Conversion Rate Optimization in e-Business: Trust, Design, and Personalization Implications for Landing-Page Design. Springer Nature Switzerland, Strategic Innovative Marketing and Tourism series. DOI: 10.1007/978-3-032-12968-0_37
- Unbounce, Conversion Benchmark Report. Approximately 41,000 landing pages, 464M pageviews, 57M conversions.
- LeadzLander, The State of SaaS Landing Pages (2026). 800 attempted, 675 scored, collected 2026-07-25, seed 20260725.