A creative taxonomy is a short list of fields you apply to every asset: what it says, what it looks like, where it ran and whether it is still safe to use. Seven fields cover reuse for most teams, and they give an AI tool enough to filter assets reliably. Any more than that and tagging turns into a chore nobody finishes.
TLDR: Tag each asset with seven fields (topic, format, hook type, brand, rights end date, status, best result). Use one naming pattern, one owner for the list of allowed values, and a quarterly clean-up. The point is deciding in seconds whether to reuse, adapt or retire an asset.
Tag for the reuse decision, not for filing
Most libraries are organised like filing cabinets: by year, by campaign, by team. That helps you find where something was stored. It does not help you answer the question you have on a Tuesday: "Do we already have a short video about this product that we can use again?"
So choose fields that answer reuse questions. This post covers asset-level fields for creative. For the content-level view, see the six-field content taxonomy, and for folders and naming, the multi-brand media taxonomy guide.
The seven fields
| Field | Allowed values (example) | The question it answers |
|---|---|---|
| Topic | Product, customer story, how-to, event | What is it about? |
| Format | Carousel, short video, single image, text card | Can it fit this channel? |
| Hook type | Question, number, claim, demo | What opens it? |
| Brand | One value per brand | Whose voice? |
| Rights end date | A date, or "none" | Is it legally safe to reuse? |
| Status | Live, approved to reuse, retire | May we use it now? |
| Best result | Reach or engagement versus account median | Is it worth reusing? |
The rights date is the one people skip most, and the most expensive to forget. A licensed photo or a music clip that expired last month can cause real trouble when it gets reposted.
Keep the allowed values closed
If anyone can type any value, you will end up with "how-to", "howto", "How To" and "tutorial" for the same thing. Give each field a closed list of values and one person who can add to it.
Review the lists once a quarter. Merge near-duplicates and delete values nobody used.
Use a naming pattern that matches the fields
One pattern, written down: brand_topic_format_YYYYMM_version. A file called acme_howto_carousel_202610_v2 tells you most of what you need without opening it.
Names are the backup when tags get lost during an export. They also let anyone scanning a folder spot a gap.
Why the result field matters for reuse
"Best result" is how you separate assets worth reusing from the rest. Record the result as a ratio to the account's own median, not as a raw number, so results across accounts compare fairly.
In Mydrop's September 2026 data, carousels ran 10% above an account's typical post. Your own data may differ, which is why you record it for your assets and do not assume.
Where Mydrop and AI help with the library
You can keep the tags in a spreadsheet, but the value shows when the library connects to the calendar. In Mydrop, AI content tools can draft new captions and variants from a brief, and bulk creation turns a table of approved assets into scheduled posts. Workspaces keep each brand's assets and permissions separate.
Be careful with automatic tagging. Let an AI suggest the topic and format, but have a person confirm status and rights, since those two carry the risk.
Clean the library every quarter
A quarterly routine takes about an hour:
- Mark anything with an expired rights date as "retire".
- Tag the five best results of the quarter as "approved to reuse".
- Merge duplicates and fix naming drift.
- Remove tags nobody searches by.
For the next step, take your 50 most recent assets and tag them with the seven fields in a sheet. Time it. If it takes longer than two minutes per asset, drop a field. Then connect the sheet to your planning flow in Mydrop.
Google review
Trustpilot review
Product Hunt comment
