Art inventory software, written by people who have shipped to Basel

What "gallery inventory" actually is

A working gallery inventory is not a list of artworks. It is a record of every object the gallery is responsible for, plus the agreements, locations, conditions, and people attached to each one. A 30-work show is not 30 rows; it is closer to 30 objects, 30 condition reports, 12 consignment agreements, 8 storage locations, 4 shipping legs, and 2 insurance certificates, all of which need to agree on the same titles and dimensions.

When the Inventory tab is the place every other workflow reads from, things stay aligned: invoices show the same dimensions as the wall label, viewing rooms show the right edition number, and the customs broker gets the same artist name spelled the same way as the catalogue.

When inventory lives in three places (a Dropbox folder, an Excel sheet, and the director's memory), all three eventually disagree, and you find out at the worst possible moment.

The five things every gallery inventory has to record

Galleries that scale past 100 works without pain are the ones that decide, on day one, that these five things live in the inventory record itself, not in a Word document or a separate folder.

1. Identity. Artist (a stable record, not a free-text string), title, year, medium, dimensions in centimetres and inches, edition info if any. Every other field cascades off of these.
2. Photos. A primary install image and at least one detail. Filenames that mean something six months later, not IMG_4821.jpeg. The same files that get sent to fairs, consignors, and insurers.
3. Status. Available, reserved (for whom, until when), on hold, sold, returned, destroyed. Three values is not enough; ten is too many. Pick five and stick to them.
4. Custody. Where the work physically is right now (storage, gallery floor, framer, in-transit, on loan to a museum). Each move is a dated event, not a field someone overwrote.
5. Commercial terms. Owner, consignor split, retail price in the gallery's primary currency, any pre-agreed discount ceiling. This is the field that prevents the most expensive arguments.

Where Excel, Airtable, and Dropbox break first

Almost every gallery starts with a shared Excel sheet, photos in Dropbox, contracts in Google Drive, and contacts in Mailchimp. This works until somewhere around the second exhibition or the 80th work, at which point three things start happening simultaneously: people stop trusting search ("is that really every Mary Smith we have?"), images and rows separate (two people uploaded the same install shot under different filenames), and consignment splits get re-asked in Slack because nobody remembers which version of the spreadsheet has them.

The breaking point is not the volume of data. It is the number of people touching it. A solo artist with 600 works in a tidy spreadsheet is fine. A 4-person gallery with 80 works in a spreadsheet is already in trouble, because they are now arguing about which row is canonical.

Airtable solves the multi-user problem and creates a worse one: you now have a flexible database with no opinions about how artworks should be modelled. Every gallery on Airtable eventually rebuilds the same five tables (works, artists, contacts, exhibitions, sales), badly, and discovers that the bit they really needed (a viewing room, an invoice, a real CRM) is two integrations away.

A workflow that survives Frieze week

The week before a fair, what changes is not the size of the inventory. What changes is the number of people who need to read from it (now including the booth team, the shipping company, the press office, and a few collectors who got early PDFs). A workflow that survives looks like this:

1. Lock the booth checklist 10 days out. After the lock, dimensions and titles do not change. Marketing copy can still change; operational fields cannot.
2. Generate the manifest from the inventory itself, not from a re-typed PDF. The customs broker reads from the same fields the wall labels do.
3. Build the private viewing rooms for early collectors directly from the booth checklist, so when a work moves from "available" to "reserved", every room reading from that work updates.
4. Reserve works against named collectors with an expiry time, not as a vague "hold for X". A reserve with no expiry is a hold forever.
5. After the fair, run the returns leg as a single batch update. Every work that came back gets a custody event the same day, before anyone goes home.

How Art.industries handles inventory

Inventory in Art.industries is the product's spine. Works, Artists, Collections, Exhibitions, Publications, and the Network (CRM) are linked records, so the artist record on a work is the same artist record on every contract, contact note, and exhibition. Edit a name once and it propagates.

Edition handling is built in. A 1/5 edition is one work record with five edition copies, each with its own status and owner. Selling 2/5 does not require duplicating the work; it updates the relevant edition copy. The catalogue, viewing rooms, and invoices all read from the same edition state.

Consignment splits, retail prices, and the sales-stack lifecycle (reserved, invoiced, paid, shipped) sit on the work itself, not in a parallel sheet. When you generate a Stripe-backed invoice from a work, the title, dimensions, edition number, and price are not retyped; they are pulled from the record. See Art invoicing software for the invoicing flow.

Custody is modelled as a series of dated location events, so "where is 1/5 of the Mary Smith?" has a definite answer (and a history). That same data feeds the multi-location inventory view used by registrars during fair season.

Common mistakes when setting up inventory the first time

Most painful migrations are caused by a small number of avoidable decisions. The five below cause about 80% of the cleanup work later.

• Treating "Mary Smith" as a string. Make every artist a real record from day one, even if you only represent four artists today. The day you decide to also collect "Smith, M." was the day search broke.
• Recording dimensions in only one unit. Centimetres for European fairs, inches for American invoices. Store both, computed from one, never re-typed.
• Skipping the edition copy model. A 1/5 edition is five sales, five owners, five custody chains. Modelling it as one row will hurt the second time you sell it.
• Letting "available" mean three different things. Reserved-but-not-invoiced is not the same as on-hold-for-the-artist is not the same as in-the-back-room. Pick the values up front and write them down for the team.
• Putting consignment splits in a Word doc. Splits live with the work. The Word doc lives in a folder no junior staff member knows about, until the artist asks for a statement.

Inventory for solo artists, galleries, and museums

A solo studio (see Software for artists) typically optimises for fewer fields, more photos, and fast invoicing for editions and merchandise. The inventory has one consignor (the artist), one consignment agreement (themselves), and the commercial layer is mostly about studio sales and small editions.

A commercial gallery (see Software for art galleries) sits in the middle: multi-consignor, multi-location, fair-driven, with viewing rooms and invoicing as core weekly workflows. This is the canonical fit for the system as designed.

A museum or institution (see Software for museums) leans harder on loans, condition reports, and publication history, with less weight on private sales. Art.industries handles the loan and condition side, but if you are running a 50,000-object permanent collection with deep accession provenance, you are probably evaluating us alongside a dedicated collections management system.