Twenty years of collections, in one place

Northbank had been collecting since 2004. Twenty-one years of ceramics, weaving, glass, metalwork and stone, catalogued across four systems that had each been the obvious choice at the time and none of which spoke to any of the others. There were index cards. There were three spreadsheets nobody would admit to owning. There was a database on a machine under a desk that the technician had been asked, repeatedly, not to turn off.

The brief we were given was to build an archive site. The brief we ended up working to was different, and we did not arrive at it quickly.

What the first month was for

We spent the first four weeks not designing. We sat with the two curators and the collections assistant and watched them answer questions, because the questions people actually asked turned out to be the whole shape of the problem. Almost nobody arrived wanting to browse twenty-one years of objects. They arrived wanting one thing: whether Northbank held anything by a particular maker, or in a particular material, or from a particular year, and whether they could come and see it.

That is three questions, and the old system answered none of them without a person in the loop. Every enquiry became an email, and every email became a search across four systems by someone who already knew where things were. The institute was not short of a website. It was short of a way to let its own material answer for itself.

So the structure we built is narrower than an archive and more useful than a catalogue. Three ways in, and nothing else on the front page:

• By maker, because that is how most people already think about craft.

• By material, because that is how the collection is physically stored and how curators think about it.

• By year, because exhibitions and loans are almost always organised around a period.

Everything else — the essays, the exhibition history, the acquisition notes — sits behind those three doors rather than competing with them. A visitor who wants to wander can still wander. A visitor who wants one object can reach it in two clicks.

Building it so they can keep it

The part of this project we are most pleased with is the part that is hardest to show in a photograph. Northbank has no developer. It has two curators, an assistant, and a technician who maintains kilns rather than servers. Anything we built that needed us to maintain it would have quietly decayed the moment our involvement ended, and we have seen enough institutional websites to know exactly how that goes.

So every field a curator touches is plain language, and there are fewer of them than they expected. Object title. Maker. Material. Year. Dimensions. Three images. A description in ordinary sentences. That is the whole record. There is no taxonomy to maintain, no controlled vocabulary to police, no relationship graph to keep consistent. Materials and makers are typed as text and grouped automatically, which means a typo produces a visible duplicate the curator can fix in ten seconds rather than an invisible orphan they will never find.

If you get one thing right, make it the search. Nobody comes to an archive to admire the archive.

We wrote the editing guide as eleven pages with screenshots, and then we did the thing that actually mattered, which was to sit with the collections assistant while she added forty objects herself, and fix the seven places where the interface had assumed knowledge she had no reason to have. Six of those seven fixes were wording. One was a genuine design mistake on our part, in the image ordering, and it would never have surfaced in a review among designers.

Where it stands

Fourteen hundred objects are catalogued and photographed. The site is in English, Welsh and Polish, the three languages the institute’s audience actually reads, with the translation workflow handled inside the same interface rather than as a separate export. The four old systems are decommissioned. The machine under the desk has been turned off, with ceremony.

There is a second phase we have deliberately not started. Northbank would like to open the archive to contributions from the makers themselves, and that is a different project with different risks, most of them about moderation rather than design. It will be easier to do well in a year, when the curators have lived with the current system long enough to know what they actually want from the next one. We have written down what we would change, and left it with them rather than keeping it as leverage for a second engagement.