Most websites are designed like a building. You come in the front door, read the directory, and take the hallway to the room you need. The homepage is the lobby, the navigation is the directory, and the whole thing assumes people walk in the same way.
At Global Fishing Watch we had that assumption too. Then we checked it.
How we asked
Nothing fancy. We combined three sources.
Analytics told us where sessions started and where they went next.
Heat maps showed what people clicked on the pages they did see, and what they scrolled past.
A survey asked people directly why they came. It was a simple Google Form. We promoted it with a popup shown to 5% of site visitors, in the organization's email newsletter, and on social media. The popup reached people in the middle of a real visit. The newsletter and social posts reached people who already know us well. Responses are still coming in, so the survey is confirming patterns rather than setting them.
None of these is conclusive alone. Analytics tells you what happened but not why. A survey tells you why but only from the people willing to answer. Where all three agreed, we treated it as true.
What we found
Over 90% of visitors never visit the homepage. They arrive directly on editorial articles, usually from search or a shared link. For most of our audience, the article is the website.
Over 60% of homepage visitors are trying to reach the interactive map. Of the people who do come in the front door, most want to walk straight through it to one place.
So the lobby we had spent the most care on was mostly a pass-through. And the pages doing the real work of introducing people to the organization were articles, which were designed as the last stop on a trip nobody was taking.
Why the navigation was built the way it was
Our navigation was organized by "programs" and "topics." That is how the organization is structured internally, and it makes complete sense from the inside. Every team can find itself.
It makes much less sense to someone who arrived on an article about a specific fishing fleet and wants to know what else we have. They do not know our program names. They should not have to.
This is the most common navigation problem I see, and it is not a failure of anyone's effort. The org chart is the structure everyone in the room already agrees on, so it is the easiest one to build. Visitors are not in the room.
What changed
The findings ended the argument about which path users take, because the data said most of them take no path at all. That made the rest of the decisions easier.
The navigation stopped mirroring the org chart. It is now built around three things a visitor can actually use: what we do as an organization, the tools we offer, and the impact that work is having on the world.
New page templates are in development, starting from the assumption that any page can be the first page. Each one is built from three sections:
Why this matters. The stakes, stated plainly, for someone who has never heard of us.
What we do and the real world impact. Our work on the issue, and what it has changed.
Resources. Links, PDFs, and tools the visitor can take with them.
A visitor who lands on one of these pages from search gets the context the homepage used to carry, without having to find the homepage.
What I would tell anyone doing this
Check where sessions start before you redesign the homepage. It is a five-minute report, and it can move your whole budget.
Treat entry pages as front doors. If most traffic lands on articles, the article template is your most important design surface.
Ask people why they came. A free form and a popup to a small slice of traffic is enough. You do not need a research budget to find out your structure is built for you.
Use more than one source. The survey alone could have been written off as a vocal minority. The analytics alone would have raised a question without answering it. Together they ended the debate.