The problem with a startup website is not the design. It is that the company describes itself differently in six months, and the site was built as a single artefact. Norml builds the system underneath instead — so changing the story is an edit, not a rebuild.
A few clients we’ve worked with
The site was a landing page for launch, and every new audience since has been bolted onto it.
A founder writes the copy at midnight, a designer redraws a screen, and the two versions of the brand drift apart.
Investors, customers and candidates all get the same page, and it is pitched at whoever the last raise was for.
The product changed, and updating the site means opening a Figma file somebody left the company with.
Everything looks like every other startup site, because it came from the same template and the same three reference sites.
Seven things, in the order we do them. Web design for startups is a scope question before it is a design question — the goal is the smallest system that will not need replacing at the next stage.
What the company is, for whom, and why now — agreed before anything is designed, because a site built on an unsettled story gets rebuilt. This is where the InsuraTec engagement started too, and it is the cheapest hour in the project.
Three audiences arrive at an early-stage site and want different things. Which pages exist for each, and which facts are shared. Most startup sites serve the audience that mattered at the last fundraise and quietly lose the other two.
Enough identity to be recognisable and distinct from the twenty companies pitching the same round, without the cost of a full brand programme. InheriNext and GoPetZen were both built this way — brand and site together, from nothing.
Components, states and responsive rules in Figma — but scoped to what you need now, with room to grow. A founder or a junior designer should be able to make a new page that looks right without asking anyone.
Campaign, launch and experiment pages built from the same system, so testing a new message costs an afternoon rather than a design sprint. This is usually the capability that pays for the engagement.
WordPress where the team is non-technical and the content moves, Next.js where the site has to sit next to a product. We pick the one you can actually run, not the one that photographs well.
Search Console, analytics and conversion tracking configured at launch, because at this stage the site is an instrument as much as a shopfront and you need to know which message worked.
The scope follows the stage and how much of the story is settled. A typical engagement hands over the following:
A positioning summary the whole team can repeat, agreed before design starts.
A page map covering every URL, the audience it serves, and what it has to prove.
A brand identity sized to the stage — enough to be distinct, not a full programme.
A Figma design system with components, states and responsive rules.
Page templates you can build new pages from without a designer.
A built site on WordPress or Next.js your team edits alone.
Analytics and Search Console configured from day one.
A handover session and every account and file in your name.
The services a young company usually buys together, sized so the system outlasts the current version of the story.
Our process
A short, sharp session on what the company is and who it is for. At this stage the diagnosis is usually that three people describe the product three ways.
[ 01 ]The audience split, the page map and the URL plan. Deliberately small — the goal is the fewest pages that do the job.
[ 02 ]Wireframes, then brand and the system in Figma. Components and states, sized so a non-designer can extend them.
[ 03 ]WordPress or Next.js, structured so your team ships a new page without calling us.
[ 04 ]Analytics, Search Console, speed, then the files and the session that make you independent.
[ 05 ]
