A startup website needs five things before launch: a homepage that says what you do and who it’s for, a page for each thing you sell, proof that you’ve done it, a short page about the people behind it, and one obvious way to get in touch. That’s roughly five pages. Everything else can wait, and I’d take a small site that’s sharp over a big one that says nothing.
Which pages does a startup website need on day one?
Start with the homepage. A stranger should know what you do and who it’s for before they scroll. If they have to hunt for it, the page has failed, however nice it looks.
Then one page for each main thing you sell. If you sell one thing, that’s one page. After that comes proof: a case study, a product walkthrough, whatever shows the thing working. Add an about page so people can see who’s behind it, and a contact or booking page so they have somewhere to go.
That’s the site. The website design page shows how we work through it on a real project.
What job should each page do?
One. This is where most startup sites go wrong. Every page tries to explain the whole company, so every page is a bit vague, and nothing tells the visitor what to do next.
Before I touch Figma on a new project, I slow down and ask questions. Lots of them. These four come up every time:
- 01What problem are we really solving here?
- 02What do we want people to feel when they first land?
- 03What does success look like for the product and for the business?
- 04What stories are we trying to tell, and what are we trying to avoid?
Most design problems don’t start in Figma. They start in unclear thinking. Once you know the one job a page has, the layout mostly falls out of it, and so does the call to action. Sometimes the whole difference is moving that button 50px higher so someone actually sees it.
How do I know what to say first?
Work out what your visitor is really asking. On a discovery call, the question under every other question is "do you actually understand what I’m trying to build?" A website has the same problem, and it has to answer it fast.
A founder had sent over a brief for a website project. Before we spoke, I put the brief, his competitors, his audience and the outcome he wanted into Claude. I wasn’t asking it to do the work, just to help me think faster. What came back was a positioning angle, a structure for his audience and a direction for how the thing could feel. Not final, not pixel-perfect, but specific.
So instead of asking him what he wanted, I showed him. He asked a few questions and signed before the call ended. The uncertainty had gone. I think a good homepage does the same thing: it removes the doubt in the first few seconds, so the visitor can stop wondering and start deciding.
What can wait until after launch?
A lot, honestly. The blog can wait. So can a resource hub, a careers page, a press page, and pages for features you haven’t built yet.
What shouldn’t wait is the ability to add them later. On a few sites we’ve built, the point was never launch day. It was a CMS the team could run themselves, so the site could grow with the business. One was a filterable portfolio the client’s own people update. Another was a site for a college coach, built to turn parent interest into booked consultations.
If you’re deciding how to build it, Webflow vs Framer compares the two platforms a lot of startups weigh up.
Do I need the trendy platform to launch?
No. I get asked "why not Webflow for every project?" a lot, and the truth is the best tool is the one that fits the client.
On one redesign, the team already had a WordPress setup that worked for them. So I designed a system in Figma that could be adapted straight into that environment. No bloat, no unnecessary migration, just clear design translated into a platform they could manage easily. For me that’s what good design is: not forcing shiny tools onto people, but meeting them where they are.
How long should it take?
For us, a website design runs about four to six weeks to go live, and a small scope is what keeps it there. Long timelines usually come from adding pages, and adding pages usually comes from not having decided what the site is for. More on this in how long a website redesign takes.
Founder of Rvysion. Has led brand refreshes and new identities for Honeycoin, Voxtell, Lateral Frontiers and Idara.