How to build a lead generation website that actually converts

Most website projects run in the wrong order. Design first, content squeezed in afterwards, SEO considered near launch, measurement bolted on when someone asks how it is performing.

This is the order that works. Nothing here is exotic; the sequence is the valuable part.

Step 1: find out what people actually search for

Before deciding a single page, find out what language your buyers use. It is frequently not the language you use internally.

Businesses describe themselves in terms of what they provide. Customers search in terms of the problem they have, and they use the common term rather than the correct one. If your industry calls it "thermal envelope remediation" and customers search for "damp proofing", you need a page targeting the second regardless of what you put on your invoices.

What you want from this exercise:

• The phrases with real search volume that describe what you sell
• How competitive each one is, so you know which are realistic this year
• The specific, qualified versions of broad terms, which usually have far less competition and much better intent
• The questions people ask around each topic, which becomes your content plan

The general pattern: broad category terms are dominated by large established sites and are not winnable quickly for a new or small domain. The more specific version of the same term often has meaningful volume and almost no direct competition, because nobody targeted it deliberately. Those are where a smaller business should start.

Step 2: decide the page structure before the design

One page per thing people search for and buy. That is the rule that does most of the work.

If you offer three services, that is three pages, each making its own argument and targeting its own phrases. Do not combine them because it feels tidier. A page trying to rank for three different intents ranks for none, and a visitor interested in one has to wade through two others.

Practical constraints worth keeping:

• Everything important within three clicks of the homepage
• Consistent URL patterns, lowercase, hyphenated, no dates, no changing them later
• A clear parent and child relationship, so a services hub links down to each service and each service links back up
• Related pages linking to each other where a reader would genuinely want the other page

Internal linking is doing two jobs here: helping a search engine understand which pages matter and how they relate, and helping a reader who is not ready for your contact form find the next useful thing.

Step 3: write the page before you design it

Designing around placeholder text produces layouts that content has to be trimmed to fit, and the first thing cut is always the specific detail that would have done the persuading.

A structure that works for a service page:

1. What this is and who it is for, in plain terms, at the top.
2. The problem you solve, described specifically enough that the right reader recognises their own situation.
3. What you actually do, broken into the sub-services people search for, each with a real explanation rather than a list of nouns.
4. Why you, with reasons that a competitor could not copy verbatim.
5. Evidence, positioned next to the claims it supports.
6. The objections, answered directly. Cost, timescales, commitment, what happens if it goes wrong.
7. The next step, specific and proportionate.

On length: write what the subject needs. A considered purchase supports a long page, because someone genuinely evaluating you will read it. Padding to hit a word count is obvious and it dilutes the page. Every section should earn its place by answering something.

Write the headings so they make sense read on their own. Someone scanning reads headings only, and increasingly so do AI answer engines summarising your page. Clear, specific headings that state the point serve both.

Step 4: design for reading, not for looking

The design should make the argument easy to follow. Some things that consistently help:

• A measure of around sixty to seventy-five characters per line for body text
• Generous line height and real space between sections, so the page has a rhythm
• One clear visual hierarchy, so the eye knows where to go next
• The primary action visible without scrolling on every page, and repeated at the bottom of long ones
• Diagrams instead of stock photography where the subject is abstract

Resist the carousel at the top of the homepage. It solves an internal argument about which message is most important and costs you the one thing you had, which was a clear first impression.

Step 5: make the site fast

Speed affects both ranking and behaviour, and it is far easier to build in than to retrofit.

The things that matter most in practice:

• Server-rendered or statically generated HTML. A site whose content only appears after JavaScript runs is slower for people and harder for crawlers.
• Images in modern formats, correctly sized, lazy-loaded below the fold. Oversized images remain the single most common cause of a slow page.
• Self-hosted, subset fonts. Loading five weights from a third party costs you a connection and a render delay. Two subset weights is usually plenty.
• Reserved space for anything that loads late, so the page does not jump while someone is reading.
• Restraint with third-party scripts. Each chat widget, tracker and embed has a cost, and they accumulate quietly.

Step 6: forms that people finish

The form is where the enquiry either happens or does not, and it is usually the least considered part of the site.

• Ask for the minimum. Every field costs completions. Name, email and a message is enough to start a conversation. You can ask the rest when they reply.
• Do not ask for a phone number unless you need it. It is one of the most common reasons people abandon a business enquiry form.
• Label fields properly, above the input, and mark what is optional.
• Say what happens next. "We reply within one working day" beside the button removes a real hesitation.
• Make the confirmation clear, so nobody submits twice wondering if it worked.
• Offer an alternative. Some people will not use a form. A visible email address costs nothing.
• Test it monthly. Broken forms are common and invisible, and they can lose months of enquiries before anyone notices.

Step 7: measurement, from day one

Set this up before launch, not when someone asks for a report.

The minimum worth having:

• Form submissions tracked as conversion events, with the form and page recorded
• Search Console connected, so you can see actual queries and any indexing problems
• A sitemap submitted, and a check that your important pages are actually indexed
• Some way of seeing where people stop on long pages

One caution about attribution: the report will over-credit whatever touch came last. Someone who found you through search three weeks ago, read two pages, then returned directly and enquired will often be recorded as direct traffic. Treat the numbers as directional rather than exact, and be suspicious of any conclusion that would have you cut the channel that introduces people to you.

Step 8: then improve it

Launch is the start. The first two or three months of real behaviour will tell you more than any amount of pre-launch opinion.

Where to look first:

• Pages with traffic and no enquiries. The clearest opportunity: people are arriving and leaving. Usually the next step is missing, unclear, or wrong for how ready they are.
• Pages ranking just below the first page. Improving something already close is far cheaper than starting a new page from nothing.
• The queries you are getting impressions for but not clicks. Often a title and description problem rather than a content one, and quick to fix.
• Form abandonment. If people start and do not finish, the form is asking too much.

Change one significant thing at a time where you can, and give it long enough to mean something. At the traffic levels most SME sites operate at, a fortnight of data is noise. A quarter is a signal.