Wren Archive

The Wren holds ninety thousand prints and had, until recently, a search box. Search boxes are excellent for people who already know a photographer's name and useless for everyone else, which is roughly ninety per cent of the institute's web traffic according to its own logs.

We rebuilt the archive around browsing rather than querying. The entry point is not a field but a set of oblique entrances — by decade, by process, by what is physically in the picture, by the size of the print — each of which lands the visitor in a working set of a few hundred images rather than in a result count. Search still exists, and it is very good, but it is no longer the front door.

The interface's main job is to get out of the way of the prints, so it is almost entirely typographic and almost entirely still. Images load at their own aspect ratio on a neutral ground, captions sit below rather than over, and nothing animates on scroll. The only motion in the whole system is the transition between a working set and a single print, which is deliberately slow enough to keep the visitor oriented about where they came from.

The hardest problem was the catalogue data. Ninety years of cataloguing conventions meant the same process was recorded eleven different ways, so a third of the engagement went into a mapping exercise nobody had budgeted for. The lesson, which we now assume by default, is that an archive interface project is a data project wearing a design brief.

Visits are up, but the number the institute cares about is dwell: median session length has roughly tripled, and the most-viewed prints are no longer the five that used to be on the homepage.