Every decision an institution makes about an object — whether to lend it, conserve it, insure it or exhibit it — depends on the quality of its record. Yet in many collections, documentation has grown in layers: accession registers from one decade, card catalogues from another, spreadsheets maintained by individual curators, and photographs stored on whichever drive was available at the time. The result is a record set that works only as long as the people who built it are still in the building.

Good documentation practice is less about any single tool and more about a small number of habits applied consistently. Below are the practices we see making the biggest difference.

1. Start from a documented procedure, not a template

Before deciding which fields to fill in, decide which processes you are documenting: acquisition, object entry, location and movement control, cataloguing, condition checking, loans in and out, and deaccession. Frameworks such as SPECTRUM, the UK collections management standard published by Collections Trust, are useful here because they describe each procedure and the minimum information it should produce. Even if you do not adopt a framework formally, writing down who records what, and when removes most of the ambiguity that causes inconsistent records.

2. Define a minimum record — and enforce it

A complete catalogue entry may contain dozens of fields, but a reliable inventory needs only a handful that are always present. A practical minimum record typically includes:

• A unique, permanent object number that is physically marked on or attached to the object
• Object name and a brief description
• Current location and the date that location was last verified
• Acquisition method, date and source
• At least one identification photograph
• Current condition summary

When these fields are mandatory at the point of entry, every object can at least be found, identified and accounted for — the foundation for audits, insurance and loans.

3. Use controlled vocabularies

Free text is where consistency breaks down. One cataloguer writes “oil on canvas”, another “canvas, oil”, a third “oil/canvas” — and searches start missing objects. Controlled lists for object types, materials, techniques, places and people solve this. Established thesauri such as the Getty Art & Architecture Thesaurus (AAT) and the Union List of Artist Names (ULAN) give you a ready-made starting point, and local terms can be added where international vocabularies fall short.

4. Separate facts from interpretation

Attributions, dates and provenance often change as research progresses. Record the current interpretation, but keep the history: who made the attribution, when, and on what basis. A record that silently overwrites earlier information loses exactly the evidence a researcher or a lending institution will later ask for.

5. Plan for interoperability from day one

Records rarely stay inside one system forever. They are shared with aggregators, published on public portals, exchanged with partner institutions and migrated when software changes. Structuring data so it can be mapped to common exchange formats — such as Dublin Core for basic description, LIDO for museum objects, or the CIDOC CRM reference model — protects the investment you make in cataloguing today.

6. Treat images as part of the record

Photographs should be linked to the object record, not stored alongside it. Each image needs its own minimal metadata: what it shows, when it was taken, by whom, and under what rights. Keep a high-resolution master file and generate smaller derivatives for daily use, so that the master is never edited or overwritten.

7. Make location control routine

Most “missing” objects are not lost; they are simply not where the record says they are. Record every movement at the time it happens, and run periodic spot-checks that compare a random sample of records against shelves. Spot-check results are also a simple, honest indicator of documentation quality that can be reported to management or funders.

8. Keep an audit trail

Who changed a record, when, and what the previous value was should always be recoverable. An audit trail protects the institution during inspections and disputes, and it makes role-based permissions meaningful: staff can be given broad editing rights precisely because every change is traceable.

Where to begin

If your documentation backlog feels overwhelming, start small: agree the minimum record, apply it to every new object from today, and work through the existing collection by priority — objects on display, objects going out on loan, and high-value or high-risk items first. Consistency going forward is worth more than perfection in a single pass.

Archive for Art is built around these practices, with configurable metadata fields, labels and groupings, linked media and a full record history. If you would like to see how it fits your current documentation workflow, request a demo.