Which IPTC Fields Should You Use for a Photo Archive?

Metadata

Why narrow the list

The IPTC Photo Metadata standard has dozens of fields. Trying to fill them all guarantees filling none properly: the campaign runs out of steam, and the collection ends up half-described inconsistently.

Eight fields filled everywhere beats forty filled a third of the way. An unevenly filled field is nearly unusable, because filtering on it returns a wrong answer without saying so.

The eight fields that carry the essentials

1. Creator — the author. The single most important field in a photographic archive, and the one most often missing from older collections.
2. Copyright Notice — the rights statement. Without it, a circulating image becomes orphaned within a generation.
3. Credit Line — how the attribution should be printed. Distinct from creator, and expected by publishers.
4. Description — the caption. What the image shows, in one factual sentence. This is the field documentary search engines will read.
5. Keywords — drawn from a controlled vocabulary. Entered freely, they create more noise than recall.
6. Date Created — the date the image was created, distinct from the file date. For a 1962 negative scan, 1962 is what counts.
7. Location — the place, broken into sub-fields (city, region, country) rather than one string.
8. Rights Usage Terms — usage conditions. This is what lets you answer a reuse request ten years later without an investigation.

The order to fill them in

Handle what is universal and automatable first: Creator, Copyright Notice, and Credit Line are identical across a whole collection or project. Bulk editing sets them in one operation.

Then what is per batch: Date Created and Location, usually shared across an entire shoot.

Last, what is per image: Description and Keywords, the only fields that genuinely require looking at each photograph.

This order matters. It gives you a complete baseline in minutes, then refines — the reverse produces ten perfect records and 40,000 empty ones.

Traps worth knowing

• Date Created versus file date. A digitised collection filled in carelessly ends up entirely dated to the year of scanning.
• Duplicated legacy fields. Some notions exist in two versions in the standard, and applications don’t always write the same one. Check what yours actually writes.
• Free vocabulary. Without a controlled list, portrait, Portrait, and portraits become three keywords.

In practice

1. Set Creator, Copyright, and Credit across the whole collection first.
2. Fill Date Created with the creation date, not the file date.
3. Break location into sub-fields.
4. Keep keywords within a controlled vocabulary of reasonable size.
5. Verify that an export read back by another tool finds your values.

Obscura Flow reads and preserves IPTC blocks at import, writes by merging sidecars rather than overwriting them, and lets you set universal fields through bulk editing before refining image by image.