The rule in one line: fill all thirteen Etsy tags, and cut any tag whose words already appear in a longer tag you are keeping. Etsy weights an exact phrase match heavily but also indexes the individual words in your tags and title, so a short tag sitting inside a longer one adds no word and no phrase — it is an empty slot you have written something in. Most sellers lose three or four slots that way and a couple more to blanks, and run a listing on eight working tags out of thirteen.
That is more than a third of your search surface, thrown away before a shopper has typed anything. It is also entirely fixable in about ten minutes per listing.
How many tags does an Etsy listing get?
Etsy gives every listing thirteen tags of twenty characters each, spaces included — 260 characters in total. There is no penalty for using all thirteen and no bonus for restraint, so fill every slot. A tag whose words already appear in a longer tag adds nothing, because Etsy indexes those words once.
The two rules everything else rests on
First, the format. Thirteen slots, twenty characters each, spaces included. Twenty characters is less than it sounds — "personalized cat mug" is exactly twenty, and "personalized cat mom shirt" runs to twenty-six and will not fit.
Second, the matching. Etsy gives strong preference to an exact multi-word phrase match, but it also indexes the individual words in your tags and combines them with the words in your title. Your listing is not thirteen isolated phrases. It is a pool of words with some phrases weighted more heavily inside it.
Both halves matter. The phrase half is why a tag should read like something a person would type, in the order they would type it. The word half is why a tag can be entirely redundant even though you never wrote it twice. Tags are only the eligibility stage of a two-stage system, and it is worth knowing what happens after that — how Etsy search ranks a listing covers the conversion signals that decide your position once you are in the running.
The subset problem
Suppose your tag set opens like this, for a sweatshirt aimed at cat owners.
| Tag | Characters | Verdict | Why |
|---|---|---|---|
| cat mom sweatshirt | 18 | Keep | The core phrase. Real demand, matches the product. |
| cat mom | 7 | Cut | Sits inside the tag above, word for word and in order. Adds no new word and no new phrase. |
| cat | 3 | Cut | Already indexed from the first tag. Alone, it is an unwinnable pool anyway. |
| mom | 3 | Cut | Same problem, same size of pool. |
| sweatshirt | 10 | Cut | Same again. You are competing against every sweatshirt on the platform. |
| cat lover gift | 14 | Keep | Introduces "lover" and "gift", and is a phrase people type. |
Four tags cut from the first six. That is not an unusual result, and it is the specific failure this article is named after. Every one of those four felt like it was doing something, because each contains a word that matters. None added a word or a phrase the set did not already have.
The test is mechanical. For each tag, ask whether it contributes at least one word that appears nowhere else in your tags or title, or whether it is a distinct phrase a shopper would plausibly type as written. If the answer to both is no, that slot is empty and you have not noticed.
The replacements should introduce genuinely new territory: a different product word, a different occasion, a different way of describing the same buyer. Something like crazy cat lady gift, cat owner christmas, funny cat sweater and gift for cat mom. The fastest way to find candidates like those is to pull the related keywords around a seed phrase rather than to sit and think of them.
Single-word tags are almost always a wasted slot
The reasoning above disposes of most of them, but there is one exception worth stating.
A single-word tag earns its place only when the word is itself a complete search that people perform, and no longer tag in your set already contains it. That is rare, but it happens: aesthetic and subculture words like cottagecore, established niche terms, and occasionally a proper noun. Somebody genuinely types those on their own and browses the results.
Everything else — shirt, gift, cute, funny, mom — is a word inside other people's searches rather than a search of its own. Tagging it puts you in a pool of hundreds of thousands of listings where you will never surface, using a slot that could have put you in a pool of two thousand where you might.
Plural and singular are two different queries
Etsy does some stemming, so a listing tagged cat shirt is not invisible to a search for cat shirts. But the two phrases are separate demand pools with different volumes and often noticeably different competition, so treating them as interchangeable throws away information you could have had for free.
The practical approach is to look up both and take the larger, rather than assuming. It usually is not the one you would have guessed, and it varies by category in ways with no obvious logic — some things people search singular, some plural, and the ratio is not stable across niches. Spending two of your thirteen slots on the singular and plural of the same phrase is defensible only when both have real volume and you have nothing better to put there. Usually you have something better.
The same caution applies to near-identical word orders. Cat mom shirt and shirt for cat mom are different phrases and will match differently, but they share every meaningful word, so the second is buying you a phrase match and nothing else. Decide that on the volume figures rather than on a feeling that you have covered more ground.
Related: write tags in the spelling your shoppers use, not the one your school taught you. Etsy's largest market types American spellings, so personalized, jewelry and color belong in your tags even if every other word you write is British. Follow the shopper.
Fill all thirteen Etsy tags, always
There is no penalty for using every slot and no bonus for restraint. An empty tag is a search you have chosen not to appear for.
The instinct behind leaving three blank is a good one misapplied — a sense that a weak tag might dilute the strong ones. It does not work that way. Tags do not compete with each other for weight within your listing. Each is an additional route into a candidate set, and a route leading somewhere small still beats no route at all.
The genuine risk is not dilution but irrelevance. A tag describing something your product is not will bring in shoppers who bounce, and bounced views hurt the conversion signal that decides your ranking.
One caveat worth holding onto: traffic is only half the equation. Thirteen perfect tags on a product that nets twenty-nine cents a sale is a lot of work for very little, which is why it is worth knowing what print-on-demand really costs per shirt before you optimise the listing that sells it.
Source candidates, then verify them
The failure mode in tag writing is that it feels like a writing task. It is not. It is a research task with a very short output, and the difference shows up in results.
There are three honest sources of candidates. Your own product gives you the obvious ones: what it is, who it is for, what occasion it suits, what style it belongs to. The related-keyword neighbourhood gives you adjacent phrases you would not have thought of, which is where most of the useful long tail lives. And the listings currently winning the searches you care about show you the vocabulary that market actually uses, which is frequently not the vocabulary you would have used.
Then verify each candidate against real demand data. A phrase with no measurable volume is not a long-tail opportunity, it is an empty slot with words in it. And check the set as a set, not tag by tag — run the whole set through a tag checker to catch the ones over the character limit, duplicated, or redundant against another tag in the same list. That last check is what catches the subset problem, and it catches it faster than reading down your own list trying to be objective about work you just did.
A pass over a listing you have already written
- Write the thirteen current tags in a column, with the title above them.
- Mark every tag whose words all appear elsewhere in the title or another tag. Those are your dead slots.
- Mark every single-word tag that is not a standalone search in its own right. Usually all of them.
- Look up the survivors and confirm each has measurable demand. Cut the ones that do not.
- Refill the gaps with phrases that bring new words, from the related-keyword neighbourhood rather than from memory.
- Check the finished set for character limits and redundancy before you save.
- Leave it alone for a fortnight. Etsy needs time to form an opinion about the change.
Step seven is the one people skip, and skipping it destroys the experiment. If you rewrite the tags again on day four you will never know which version worked.
What to do next
Take one listing you already have live and run it through two tools in this order.
Paste all thirteen current tags into the Etsy tag checker. It reads them together rather than one at a time, and flags every tag that breaks the twenty-character limit or duplicates words already covered elsewhere in the set. That gives you your list of dead slots in one pass.
Then take each replacement candidate through the Etsy keyword research tool before you commit it. You are checking two things: that the phrase has measurable search volume, and that the competition is something you can realistically place in. Verify, do not guess — guessing is how the four wasted slots got there in the first place.