What four marketing sites taught us about picking one conversion and cutting everything that does not serve it.
We build products, and then we build the websites that have to sell them. Over the past two years that meant marketing sites for Fixie, Jobr, Tinrate and EONLOG. Four very different companies, and every one of those sites got better the moment we answered a single question: what is the one thing a visitor should do here?
Most websites never answer it. They try to be a brochure, a pitch deck, a documentation portal and a recruiting page at once, and they end up being a place where visitors read for a while and then leave. A marketing site is not the product. It is a door. A door has one job.
For Fixie, our own goal-tracking app, the answer was blunt: the store download. So the site is one page. It explains what the app does, from daily habits to yearly ambitions with Spotify-Wrapped-style look-backs, shows the store badges, and gets out of the way. There is no blog, no about page, no newsletter popup. Every section either moves someone toward the download or it was cut.
Being our own client made that easier to admit. No approval rounds, no stakeholder who wants their department represented. Just the honest version of the question: would this section make us tap install? Most sections do not.
Jobr made it harder. jobr.app has two visitors with nothing in common: candidates who should end up in the App Store, and employers who should end up posting vacancies. The tempting move is one page that speaks to both. The result of that move is a page that speaks to neither, full of sentences carefully worded to work for two readers and landing with neither.
So we split the flows. Candidates get a path that says job hunting can feel like swiping instead of paperwork, and ends at store badges. Employers get pricing, proof and a way in. One site, two doors, one conversion each. The rule survives, it just applies per audience.
Tinrate stretched the rule furthest. A marketplace for paid expert consultations cannot convert a sceptical first-time visitor with a button. The job of tinrate.com is to earn the first click into the booking flow, and that is a trust problem before it is a design problem. So the site shows real experts with live rates and real reviews, not stock photos and promises. The FAQ is deliberately long because it handles objections before support has to. That FAQ is the least glamorous page on the site and probably the hardest working one.
The lesson: sometimes the one job is not “click here” but “believe us enough to click here”. You still name it, and you still cut everything that does not serve it.
We are builders, and the professional hazard of builders is over-building. The website is the most tempting place to do it, because it is public and it is ours. Animations, clever scroll effects, pages nobody asked for. We have caught ourselves adding all of these, and the test that stops us is the same every time: does this move the visitor toward the one job, or does it just show off?
EONLOG, the family office behind our studio, wanted the opposite of the usual investment site: say plainly who they are, how they invest, and what founders get after the close. Short pages that respect a founder’s time. It was a reminder that restraint is a feature, and that the hardest thing to ship is less.
So before your next website project, write one sentence: the visitor is X, and this site exists to get them to do Y. If you cannot write that sentence, the site is not underdesigned. It is undecided.
We build websites, software and companies. Tell us which one you need.