What Actually Makes a Website Convert?

A website converts when it loads fast, says one clear thing in the first five seconds, backs it with specific proof, and asks for one easy action. Almost everything else is decoration. The hero section alone accounts for roughly 70% of a page's conversion outcome (roast.page, 2026), so the biggest wins come from getting the top of the page right, not from adding more sections.

That is the short answer. Below is what the data actually shows, and how we use it when we build.

What does "convert" actually mean?

Converting means a visitor does the thing you want: books a call, sends an enquiry, requests a quote. A typical B2B site converts somewhere between 2% and 4% of visitors, with strong performers reaching 6–10% (Digital Applied, 2026). Knowing your real baseline matters. A services or agency page sits at a median of about 3.6%, while a lead or quote request page can reach 6.8% (Digital Applied, 2026). Chasing a number borrowed from someone else's funnel just leads to redesigns that were never realistic.

How much does the first screen matter?

More than anything else on the page. Visitors form a visual judgement in about 50 milliseconds (Nielsen Norman Group), and decide whether to stay or leave within five seconds (Lindgaard et al., 2006). Yet only 14% of pages pass a strict five-second test, where testers can correctly say what the product is, who it is for, and what to do next (roast.page, 2026).

So the hero has to be clear before it is clever. The patterns that test best are simple:

• A single, specific number as the headline lifts conversion +18% over a standard image hero (Digital Applied, 2026).
• Headlines with specific numbers ("Saves 12 hours a week") convert 22% higher than vague ones ("Saves time") (Unbounce, 2024).
• Naming the audience — "For [audience]" framing — lifts trust scores 23% (roast.page, 2026).

And what to avoid: a generic stock photo hero is the worst pattern tested, at −11% (Digital Applied, 2026). A feature-focused headline costs 24% versus an outcome-focused one (roast.page, 2026), yet 61% of pages still lead with the feature. This is exactly the kind of gap we design around — clarity first, then proof.

What kind of proof actually works?

Specific proof. Vague proof barely moves the needle. A named-customer count ("used by 8 of the Fortune 50") lifts conversion +22%, while a generic logo strip with no context manages only +8% (Digital Applied, 2026). An unattributed testimonial is worth almost nothing, at +3% — it reads as fabricated.

The wider picture supports this: 98% of buyers read reviews at least sometimes, and 71% won't consider a business rated below three stars (ProveSource, 2026). Text-based proof also holds attention longest, around 1.06 seconds of fixation versus a glance at a logo (CXL). The rule we follow: every claim names a real number, a real source, or a real customer.

How fast is fast enough?

Fast enough that nobody notices. Page speed has a direct, continuous effect on conversion. Pages loading in under one second convert at 4.4%; at four seconds and beyond, that collapses to 1.7% (Portent). Around 53% of mobile users leave a page that takes longer than three seconds (Google, 2024).

A 0.1-second improvement in load time can produce an 8.4% lift in conversions (Deloitte/Google).

We treat speed as a build rule, not a nice-to-have — it is one of the few levers that improves every section at once. (It is also a big reason we hand-code every website rather than stacking plugins.)

Why do short forms convert so much better?

Because every extra field costs you leads. A three-field form converts at about 25%; push it to seven or more fields and conversion crashes to 3.1% (HubSpot; FoundryCRO, 2026). Three fields is the sweet spot — nearly as high as a single email field, but with far better lead quality.

Phone number fields are the worst offender: making the phone optional drops form abandonment from 39% to 4% (Zuko Analytics). One classic test cut a form from 11 fields to 4 and lifted conversions 120% (MarketingExperiments). Ask for the minimum on the page; gather the rest later.

What about the call to action?

One clear action beats several. A sticky call-to-action that follows the visitor as they scroll lifts conversion +11%, and adding a second one above the fold only gains another 1% (Digital Applied, 2026) — it does not stack. Putting three or more buttons in the hero actively hurts, at −8%, because it causes decision paralysis.

The wording matters too. "Get a quote" beats "Contact us" by 14% on services pages, and first-person, personalised CTAs have lifted clicks dramatically in HubSpot's analysis of 330,000 CTAs (HubSpot). Pick one primary action and make it obvious everywhere.

FAQ

What is the single most important thing on a converting website? The hero — the first screen. It drives roughly 70% of the conversion outcome (roast.page, 2026). Make it say one specific thing, name who it's for, and offer one clear action.

How fast should a website load? Under two seconds, ideally under one. Conversion falls steadily as load time rises, from 4.4% under one second to 1.7% past four seconds (Portent), and most mobile users leave after three seconds (Google, 2024).

How many fields should a contact form have? Three is the sweet spot, converting at about 25% (HubSpot; FoundryCRO, 2026). Keep the phone number optional — making it required pushes abandonment from 4% up to 39% (Zuko Analytics).

Do testimonials and logos really help? Only when they're specific. A named, quantified claim lifts conversion +22%, while a vague logo strip gives +8% and an unattributed quote almost nothing (Digital Applied, 2026).

None of this is guesswork — it is what the largest conversion studies consistently show, and it is what we build to: fast, clear, conversion-tuned websites informed by exactly this research. If that's the kind of site you want, book a call and we'll talk through what would move the numbers for you.