Advice about high-converting landing pages usually arrives as a list of tactics with no evidence attached. Here is what the published research supports, what our own measurements show, and where the two disagree.
What visitors say they want
Ziakis, Papadopoulos, Antoniadis and Saprikis surveyed 184 people in 2026 across five dimensions of landing page design: performance, trust, content, design and personalization.
Trust ranked first. User reviews and ratings, guarantees and refund policies, and detailed product descriptions all scored above 78%. Personalization ranked last, with one exception: respondents aged 35 to 44 rated it higher than both younger and older groups.
Two limits worth stating. It was a convenience sample recruited over social media, and it measures stated preference rather than observed behaviour. People are unreliable narrators of their own decisions. Even discounting for that, the ranking is a strong signal, and it points the opposite way from where most tooling money goes.
What companies actually ship
We crawled 800 B2B SaaS company sites from Y Combinator's public directory on 25 July 2026 and scored 675 of them.
| Measurement | Result | n |
|---|---|---|
| Showed any social proof | 42.2% | 675 |
| Placed proof before the first CTA | 11.6% | 675 |
| Stated a price | 15.4% | 675 |
| Had no in-body CTA at all | 30.7% | 675 |
| Shipped no H1 | 8.1% | 675 |
Set that against the survey. Visitors rank trust first. Fewer than half of these pages carry any, and 88.4% of them ask for the action before supplying it.
Pricing tells the same story. Detailed information ranked among the top trust signals in the survey, and 84.6% of these companies make you click through or book a call to learn what it costs.
What converts, from a sample with outcomes
Our data stops at what companies ship. It says nothing about what converts, because we did not measure conversions.
Unbounce did. Their Conversion Benchmark Report covers around 41,000 landing pages, 464 million pageviews and 57 million conversions.
- Pages with one CTA link converted at 13.5%
- Pages with two to four converted at 11.9%
- Pages with five or more converted at 10.5%
- 68.2% of pages carried more than five CTA links; 14.8% used one
- Pages written at a 5th to 7th grade reading level converted at 11.1%, against 5.3% at a professional reading level
Their population differs from ours. Every page in their sample was built by someone who bought a conversion optimization tool. That generalises to CRO-tool buyers. Our sample generalises to startups. Use theirs for what converts and ours for what your peers are doing.
On reading level, the two datasets combine into something actionable. Unbounce found simpler copy converts at roughly double the rate. We measured a median Flesch-Kincaid grade of 10.4 across the 595 pages with enough prose to score, with the upper quartile at 12.5. Most SaaS companies write several grades above the level that performs best.
The order of operations
Based on the above, in order:
1. Ask for something. 30.7% of funded startups do not. Check the body of your page with the header and footer mentally removed.
2. Ask once. The Unbounce gradient is monotonic: fewer links, higher conversion. The median page in our sample carried 2 CTA elements and 1 distinct label, which is close to right. The p75 was 3 and 2.
3. Put proof above the ask. Only 11.6% do. A logo wall, a customer count, one specific testimonial with a name attached.
4. Say the price, or say why you cannot. "Contact us for pricing" is a cost you are passing to the visitor.
5. Write it three grades simpler than feels professional.
Things the evidence does not support
Personalization as a priority. It ranked last in Ziakis. Most companies would gain more from adding a testimonial than from a recommendation engine.
Any claim about the fold. Our crawler measures document-order position, not viewport position. Our median of 34 words before the first in-body CTA is not a statement about what is visible on a screen. Measuring the fold requires a rendered viewport at a stated size, and any source quoting fold statistics without naming the viewport is guessing.
Length rules. The median page in our sample ran 759 words, with the interquartile range from 458 to 1,218. That spread is too wide to support a target.
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.