What job is this website actually doing?
Every website has one primary job, even when it looks like it is doing several things at once. Before you think about colours, fonts, or which platform to build on, name that job in a single sentence. Almost every business site falls into one of three categories.
Lead generation. The site's job is to turn strangers into enquiries — phone calls, WhatsApp messages, quote requests, consultation bookings. Success is measured in leads, not pageviews. Every page should be quietly steering a visitor toward a form, a call button, or a chat link.
E-commerce. The site's job is to complete purchases directly, with no human in the loop. Success is measured in completed orders and average order value. Every extra click between "add to cart" and "payment confirmed" is a leak in revenue, which is why Chapter 5 treats it as its own subject.
Brand or portfolio. The site's job is credibility. It exists so that when someone already interested in you — through a referral, a search for your name, a business card — checks you out, they find something that confirms they made the right call. Direct conversions matter less here than trust and clarity.
Plenty of real sites blend two of these — a consultant might run a lead-generation homepage alongside a portfolio section — but one should still be primary. That single decision quietly decides what goes above the fold, what the homepage says first, and what number you watch after launch.
Know who you are actually building it for
"Everyone" is not an audience. A site written for everyone tends to say nothing clearly to anyone, because clear, persuasive writing requires picking a specific reader and addressing their specific situation. Before any content is written, get concrete about who is actually going to land on the site.
This does not require formal market research for most small and mid-sized businesses. A short, honest paragraph is usually enough: who they are, what problem brought them here, what they already know, what would make them hesitate, and what device they are most likely using. A local clinic's typical visitor is different in every one of those dimensions from a B2B software buyer, and the site should read differently as a result.
- What specific problem or need brings someone to this site right now?
- Have they already compared alternatives, or are they just starting to look?
- What objection or doubt is most likely to make them leave without converting?
- Are they most likely to be on a phone, mid-commute, or at a desk with time to read?
- What would make them trust this business enough to hand over contact details or a card number?
Write the answers down. They will resolve arguments later about what the homepage should say first, what goes in the navigation, and what tone the copy should take — arguments that otherwise get settled by whoever has the loudest opinion in the room.
Look at what competitors are already doing
A competitor review is not about copying — it is about establishing a baseline. Visit five to eight sites that a real prospective customer would also find, and look at them as that customer would, not as a designer. Note what information they all include (that is now a visitor expectation, not a nice-to-have), what they are missing (that is a gap worth filling), and how they talk about pricing, process, and credibility.
Pay particular attention to structure rather than visual style: how many pages do they have, what is in the main navigation, do they lead with price or with trust-building first. Screenshotting five homepages side by side usually reveals a pattern faster than any amount of discussion — most competitors converge on the same handful of decisions, and the ones that stand out are worth a second look, for better or worse.
Sketch the sitemap before anyone touches design
The single most common planning mistake is starting with a homepage design before anyone has listed every page the site actually needs. This produces sites that grow a navigation menu by accident, page by page, until it is cluttered and inconsistent — a problem Chapter 3 covers in more depth. Fix the order instead: list every page first, group them logically, and only then think about how any individual page should look.
The exact list of pages depends on the site's job, but most business sites converge on a similar skeleton:
- Lead-generation sites typically need: home, a page per service or offer, an about/credibility page, a portfolio or case-study section, a contact page with a clear form, and often a FAQ.
- E-commerce sites typically need: home, category pages, individual product pages, cart and checkout, an account area, and policy pages for shipping and returns.
- Brand and portfolio sites typically need: home, about, a portfolio or work-samples section, testimonials or press mentions, and a simple way to get in touch.
Write this list out — even as plain text in a document — before opening any design tool or website builder. It becomes the input to Chapter 2's platform decision and Chapter 3's navigation structure, and it is far cheaper to revise a bullet list than a half-built page.
Before you write a word of copy or open a website builder, write down the one number that will define whether the site worked — enquiries per month, orders per week, referral conversations started. Every design decision downstream should serve that number.
Turn the plan into a one-page brief
Consolidate everything above — the site's job, the audience description, the page list, and the success metric — into a single page. This becomes the brief you hand to a designer or developer, or simply the document you keep open while building it yourself. It is short enough to read in two minutes and specific enough to settle most disagreements that come up later about what a page should say or where a button should go.
This is also the moment to be honest about budget and timeline, because both feed directly into the platform decision in Chapter 2. A five-page lead-generation site with a modest budget and a three-week deadline points toward a very different platform choice than a fifty-product store with a development team and six months to launch. Planning without acknowledging those constraints just produces a brief nobody can actually execute.