Designing discovery for travellers who have not picked a place yet

Travel sites assume you know where you are going. The search box wants a destination, and the whole experience unfolds from that one answer. But a large share of trips start earlier than that: someone has a week in November, a rough budget and no idea where to spend it.

Touristoo is built for that earlier moment, and most of the design work went into the part before a place has been chosen.

The empty box is a bad first question

An empty search field asks for the one thing the traveller does not have yet. The usual workaround is a wall of inspirational photography, which is pleasant and does not narrow anything down — you scroll, you feel something vague, you leave and open a different tab.

What people can answer at that stage is not the destination. It is the shape of the trip.

Search by the trip, not the place

The things travellers know first are constraints and mood: how long they have, roughly what they can spend, what month it is, whether they want to move constantly or sit still, and who is coming. Those are answerable in seconds, and together they eliminate most of the world.

• Budget and season are hard constraints and belong first — they remove more options than anything else.
• Pace is the one people forget to ask about and the one that most often ruins a trip. Slow and packed are different holidays.
• Who you are travelling with changes the answer completely. A place that is wonderful alone can be miserable for a group of six.
• Distance matters less as a number than as travel time, which is what people are actually budgeting.

Filters have to be legible, not exhaustive

Every filter you add makes the interface more capable and less usable. We kept a filter only when a traveller could predict what turning it on would remove. If the effect is not obvious, people either avoid it or use it once, get a confusing result and stop trusting the whole set.

A filter nobody can predict the effect of is not a feature, it is a source of doubt about the results. Fewer, more legible controls beat a complete taxonomy.

Detail is the point, not listicles

Once someone is considering a place, ranked lists stop helping. What they need is specific: what it costs to be there, which months are actually good, what the place is like rather than how it photographs. That level of detail is expensive to produce and it is the whole reason to visit a discovery site instead of scrolling somewhere else.

Discovery and planning cannot be separate

The other failure we designed around is the handoff. People find things in one place and plan in another — usually a notes app that gets opened the day before departure. Every one of those transitions loses information, and the shortlist is where the thinking lives.

So saving is the same action as planning. Anything you save goes into the trip, and the itinerary assembles from what you saved rather than being a separate document you build later. It also means a half-finished shortlist is still useful, which matters because most trips are planned in short bursts over weeks.

What we would do differently

We underestimated how often trips are planned by more than one person. Sharing was treated as an export step near the end, when in practice the argument about where to go is the planning. Getting a shortlist in front of everyone earlier would have changed the shape of the product, and it is the part we would design differently from the start.