Landing pages: what actually converts (and what costs a lot for nothing)
25 juillet 2026
A landing page has one job
A landing page isn't a shorter brochure site. It answers one specific intent, arriving through one specific channel, and offers one action.
As soon as there are two competing calls to action, "request a quote" and "explore our services", the page becomes a menu. A menu gets browsed; it doesn't convert.
The structure that works
In order, with no detours:
- A promise in one sentence. What you do, for whom, with what outcome. Not a slogan.
- Immediate proof. Client logos, a verifiable figure, a photo of the actual work. Above the first scroll.
- The problem, in the visitor's own words. If they don't recognise themselves in the first section, the rest won't get read.
- What you actually offer. Three to five points, not fifteen.
- Objections handled. Price, timeline, commitment, what happens next. A short FAQ does the job.
- A short form. Every added field has a measurable cost.
Decorative sections between those blocks add nothing. They lengthen the path.
The form is where most is lost
A six-field form converts markedly worse than a three-field one. The practical rule: ask only for what you need to get back in touch and qualify.
Name, email and a free-text sentence are almost always enough. Phone, budget and company size can be asked at the next step, once the person is already engaged.
Two often-neglected points:
- The confirmation message. If it isn't explicit, some visitors resubmit the form or leave unsure.
- The error message. A silently failing form is acquisition budget spent for nothing, and it can run for weeks unnoticed.
Speed isn't a technical detail
On mobile, on patchy 4G, a heavy page loses visitors before it's even read. The effective levers are few:
- images at the right format and size, lazy-loaded below the fold;
- no JavaScript library imported for a single visual effect;
- fonts limited to two weights, served locally;
- statically served HTML when the content doesn't change on every visit.
A statically generated page served from a cache starts with a lead that few later optimisations make up.
Local SEO, where it applies
For a local business, most of the work comes down to a few things: a name, address and phone number identical everywhere, a complete Google Business profile, LocalBusiness structured data, and content that names the areas actually served.
There's no point creating a page per town if the content is the same. Thin pages don't rank and dilute the rest of the site.
What costs a lot and returns nothing
- Scroll animations across the whole page. They delay reading and hurt perceived performance.
- The hero carousel. Visitors see the first slide and nothing else.
- The background ambience video. A few megabytes for zero information.
- The chat widget that opens by itself after five seconds.
For the same budget, these are better replaced with sharper copy and a shorter form.
Measure, or don't bother
Without tracking form submissions and where they came from, no optimisation decision is possible. It's the first thing to set up, before the first line of content, not after launch.
A concrete example of a delivered landing page: ClimatConfortOI. Similar project? Let's talk.