Landing Page Copy That Converts: What the Conversion Data Actually Shows

What a good conversion rate actually is

Most landing page arguments are unwinnable because nobody in the room knows what normal looks like. Someone says 2% is fine. Someone else says it should be 10%. Neither has a reference point.

There is one. Unbounce's Conversion Benchmark Report analysed 41,000 landing pages, 57 million conversions and 464 million pageviews, and the headline number is a median conversion rate of 6.6% across all industries.

The median matters more than the average here. Conversion data is badly skewed — a handful of extraordinary pages drag any mean upward and leave you comparing your page against outliers.

• 6.6%median conversion rate, all industries
• 3.8%median for SaaS — the lowest
• 12.3%median for events — the highest
• 41,000landing pages in the sample

The spread is the useful part. A 5% conversion rate is mediocre for a legal services page and excellent for a SaaS trial signup. Same number, opposite conclusions.

Where attention actually goes

The "above the fold" argument is usually fought with opinions. Eye-tracking research from the Nielsen Norman Group settles most of it with numbers.

Users spend 57% of their viewing time above the fold, and 74% within the first two screenfuls. More than 42% of viewing time goes to the top 20% of a page, and more than 65% to the top 40% — regardless of how long the page is.

That last clause is the one people miss. Making the page longer does not redistribute attention downward. It just adds material almost nobody reads.

The headline has exactly one job

The headline does not have to be clever, and it does not have to describe your product. It has to confirm, in the reader's own words, that they are in the right place.

This is message match. Someone clicks an ad about reducing invoice processing time and lands on a page headlined "Intelligent Finance Transformation". The click and the page are about different things, and the reader leaves before evaluating anything else on it.

A headline earns its place when it survives three questions:

• Does it echo the promise of the click? If the ad said one thing and the page says another, nothing downstream matters.
• Is it specific enough to be falsifiable? "Grow faster" is not a claim. "Cut invoice processing from nine days to two" is.
• Could a competitor put their name on it? If yes, it is a category description, not a headline.

Landing page copy examples, before and after

The examples below are constructed for illustration — they are not client results, and no conversion figures are attached to them, because a before/after pair without a controlled test proves nothing. What they demonstrate is the shape of the edit.

Every row is the same move: replace a sentence that could appear on any competitor's page with one that could only appear on yours.

CTA optimisation: the button is the smallest part

Most CTA advice is about button colour, which is close to the least important variable on the page. The decisions that matter are these:

What the click costs

"Book a demo" asks for a calendar slot and a sales conversation. "See the workflows" asks for nothing. Those two buttons suit different levels of reader commitment, and putting the expensive one in front of a first-time visitor is the most common conversion error on B2B pages.

Whether the label describes the next screen

A button is a promise about what happens next. "Submit" describes what you do with the form. "Get the checklist" describes what the reader gets. The second one is the only one written from the reader's side.

How many different things you are asking for

A page with a demo request, a newsletter signup, a pricing link and a chat widget is not offering choice. It is asking the reader to make a decision you should have made for them.

Every form field is a price you pay

Form length is one of the few places where the cost of a design decision has been measured directly. Baymard Institute's checkout research — drawn from more than 200,000 hours of usability testing — found the average US ecommerce checkout contains 23.48 form elements, while an ideal flow can be as short as 12.

The behavioural findings are blunter: 18% of users have abandoned an order because the form was too long or complicated, and test participants became visibly intimidated by forms showing 10 to 15+ fields. Baymard's testing indicates most sites can cut displayed fields by 20 to 60%.

• 23.48form elements in the average US checkout
• 12elements in an achievable ideal flow
• 18%have abandoned over form length
• 20–60%typical achievable field reduction

That research is about checkout, and a lead form is not a checkout — but the mechanism transfers, because the mechanism is the reader deciding whether the reward justifies the effort visible on screen. Ask for what you will genuinely use in the next step, and nothing else.

Run a review, not a rewrite

The failure mode on landing pages is not bad writing. It is that nobody checks the page against a fixed standard before traffic is bought, so the first honest feedback arrives as a bounce rate.

A review is cheap and repeatable in a way that a rewrite is not. Score the page against the same rubric every time:

• Message match — does the headline echo the click that produced the visit?
• Falsifiability — is there a claim on the page that could be proven wrong?
• Attention order — is the one thing that must land inside the first screenful?
• Cost of the CTA — does the ask match how much the reader currently trusts you?
• Form cost — is every field used in the next step?
• Substitution test — could a competitor paste their logo on this page unchanged?

Running that list before spending on traffic catches the errors that A/B testing cannot fix later. A test can tell you which of two headlines wins; it cannot tell you the page was answering the wrong question all along.

If you would rather not assemble the rubric yourself, the AI Marketing Automation Vault includes a landing page analyzer and a CTA optimisation workflow among its 50 systems — each with a trigger, a data contract and a QA checklist, so the review runs the same way every time rather than depending on who is doing it. It pairs with the Sales Vault when the page's job is to feed a pipeline, and both are included in the Complete Automation Bundle.

If the page itself needs rebuilding rather than reviewing, that is what we do as a service — and if the traffic is coming from cold email, the deliverability side is covered in the Email Deliverability Toolkit, because a page that converts is worth nothing if the email never reached the inbox. On that subject, see why cold emails land in spam.