My listings spent a month in Desk Calendars
Nineteen Etsy listings, none of them a calendar, all of them filed under Desk Calendars for a month. The category is not a label. It is most of the search.
Nineteen finance templates. Gratuity calculators, a general ledger with VAT, a cash-flow forecaster, a margin analyser. Every one of them listed on Etsy, live, with decent titles and thirteen tags each.
Every one of them filed under Paper & Party Supplies › Paper › Calendars & Planners › Desk Calendars.
Not one of them is a calendar.
How it happened
I wrote the publisher to pick a category automatically. It fetched Etsy's full taxonomy tree, filtered to the leaf categories, and picked the shortest path matching a preference list that started with "Calendars & Planners". Shortest path, first match, done.
The code did exactly what I told it to. The preference list was the bug, and it was a bug of the specific kind that never throws an error: the listings published successfully, the API returned 200, the log said PUBLISHED, and nothing anywhere said this is a spreadsheet in the calendar aisle.
Then I added a new kind of product and made the same mistake in a fresh way. The new rule preferred anything containing "Templates" — and the shortest path on Etsy containing "Templates" is Weddings › Invitations & Paper › Templates. So a WhatsApp sales kit and a savings calculator went live under wedding invitations. My second automatic rule was worse than my first.
Why the category carries more weight than the tags
Tags are how a listing matches a phrase. The category is how Etsy decides which shop window the listing stands in — it drives the filters down the left side of the results, the "more like this" rows, and a large share of what the recommendation system will even consider showing.
Somebody searching gratuity calculator spreadsheet and then narrowing by category will never narrow into Desk Calendars. The listing was not ranked badly. It was ranked in a room nobody looking for it walks into.
The fix, and why I stopped being clever
I had the publisher print Etsy's entire template taxonomy into the log. It is not a long list, and reading it took about a minute:
12487 ... > Templates > Personal Finance Templates
12478 ... > Templates > Bookkeeping Templates
12476 ... > Templates > Planner Templates
12486 ... > Templates > Social Media Templates
2818 ... > Templates > Website Templates
12480 ... > Templates > Contract & Agreement Templates
Personal Finance Templates and Bookkeeping Templates had been sitting there the whole time. There was never a hard problem here — there was a fuzzy string match standing between me and a list I could have read on day one.
The rule now is a dictionary. Product kind in, category ID out, no matching, no cleverness:
- gratuity calculators, budget planner, debt payoff, the savings calculator → Personal Finance Templates
- ledger, cash forecast, commission tracker, margin analyser → Bookkeeping Templates
- the web-studio kit → Website Templates
- the launch system → Planner Templates
- the video prompt library → Social Media Templates
Twenty-three listings re-assigned in one pass. Nothing else about them changed — same titles, same tags, same images, same prices.
Three things worth taking from this
Automate the boring half, choose the important half by hand. Uploading files, resizing images, writing structured descriptions — automate all of it. Deciding which shelf your product stands on is a judgement, it happens once per product type, and it is worth sixty seconds of your attention.
A silent wrong answer is the expensive kind. Everything in my pipeline reported success. If you have a step that picks something — a category, a tag set, a price tier — make it print what it picked, in words, into a log you actually read. My publisher now writes the full category path on every run.
Go and check yours. If you sell on any marketplace and you have never deliberately opened the category tree and read it, there is a real chance you are in the wrong room. It costs a minute per listing to check and it is the cheapest fix in the entire business.
I do not know yet what the traffic difference will be. That is the next thing I will have numbers on, and I will publish them here either way.

