TeePublic Keyword Research: How to Find Keywords That Sell

How TeePublic Search Reads a Listing

TeePublic has never published its ranking formula, but the inputs sit in the upload form: title, main tag, additional tags, description. Title and tags are what search visibly responds to. Nothing says description text is matched, so treat it as buyer-facing and Google-facing.

Something then sorts whatever matched, and sales history is the obvious candidate. Tiles carry the design, title and price and no rating or sales figure — checked September 2026 — so the evidence sits on product pages: open the ten designs ranking above yours and count how many carry reviews, about five minutes at thirty seconds a page. Check on one of them that the review block belongs to that design and not to the garment or the shop. If it reads shop-wide the day you look, TeePublic exposes no per-design sales evidence at all.

Both marketplaces top out at fifteen tags, so the scarcity here is not slot count. TeePublic's own tagging guidance asks for 6 to 15 total tags and then says, in capitals, to stop when you run out of relevant words. Redbubble, checked September 2026, allows 15 tags of up to 50 characters. The difference is what each marketplace does with a weak tag: TeePublic states outright that redundant tags hurt your search results, so a near-duplicate is not a wasted slot, it is a penalty.

One tag also behaves differently from the rest. TeePublic says main tags are weighted heavily in search, and the main tag additionally files your design onto a public tag page, which hands you browsers when it is busy and nothing when it is empty. That is two jobs in one field, and it is why the main tag gets its own rubric further down.

TeePublic Grades Your Tags For You, In Colour

Before any of the measurement below, use the one piece of feedback TeePublic gives away free. After upload, your tags render either blue or grey. TeePublic's guidance is that grey tags are too specific and do not have the same power as blue ones, and that a grey tag may get indexed and turn blue over time. Its stated rule of thumb: the fewer grey tags you carry, the better your chances of being found.

Read that as a live check on everything in this article. A phrase you scored highly that lands grey is a phrase the marketplace does not yet recognise — the same finding the search-bar test below produces, only confirmed after the fact on your own listing. Log the colour beside each tag on your first ten uploads and you will calibrate the scoring faster than any amount of counting.

Do not chase blue at the cost of relevance. A broad blue tag that misses the artwork pulls shoppers who bounce, and TeePublic warns separately against tagging your way into unrelated searches. Blue tells you a phrase is recognised, not that it is yours.

Why No Tool Sells You TeePublic Search Volume

Start here: it saves money. TeePublic exposes no search-volume endpoint and publishes no query report for artists. Every "TeePublic search volume" figure that can be traced leads back to Keyword Planner ranges or Amazon autocomplete, and the products selling them name no source.

Run one test. Type something nobody searches — a rare dog breed plus a hobby — and check the suggestion holds both words. If it echoes only the breed, the box dropped your second token and the test says nothing. If it carries your whole string, the box matches something broader than real searches, so never read suggestion order as a volume ranking. Nothing at all means the phrase is not in the dictionary: base demand 1.

Presence is the signal that survives, with a caveat. If the box draws on phrases sellers typed into tag slots, a phrase there proves only that TeePublic holds it as a live tag, not that buyers type it — so confirm with a Planner range before scoring it highly.

Method 1: Mine the Search Bar

Before you publish anything, the search box is the main place TeePublic tells you about its vocabulary, so TeePublic keyword research starts there. A dropdown phrase is evidence; an invented one is a guess.

Log the trigger letter beside each phrase; those reappearing under several letters are what the dictionary surfaces most readily. Every later step writes into a named sheet column: phrase; trigger letters; design count; exact-phrase titles out of 48; tag page state; Redbubble printed count; default product type, the type most of the top twenty are set to; title phrase matched, the variant those titles actually carry; relevance 0-3; demand 1-5; competition 1-5; net. Eleven of the twelve are niche-level and filled once. Relevance is the exception — re-score it per artwork, because a phrase that is a 3 for one design is a 1 for the next.

Budget about forty minutes; the a-to-z walk alone is twenty-six captures. It beats an hour in a keyword tool because the marketplace carries these phrases: "nurse practitioner student" and "np student" are one buyer and two tags, and only the dropdown says which one TeePublic holds. If a seed returns under ten phrases, drop to the parent word and re-test the children as prefixes. Aim for 30 to 45; a niche that cannot yield fifteen fails the size test.

Method 2: Measure Competition Without a Result Count

Redbubble sometimes prints a result count; TeePublic printed none as of September 2026, which is why most competition advice assumes the wrong marketplace. If a total is sitting on the results view the day you look, use it and keep the same 120 and 500 bands below. Otherwise four countable signals replace the missing number.

Count designs, not pages — tiles per page and grid width shift with the window. Where the view loads on scroll, count rows: at four across, 120 designs is 30 rows and 500 is 125.

Competition is 5 minus the Open readings: four Open scores 1, three scores 2, two scores 3, one scores 4, none scores 5. A single Saturated reading stops a phrase scoring below 3. If Redbubble prints no count, score that row non-Open rather than marking out of three.

Treat the first two bands as starting points and recalibrate after twenty phrases. Redbubble's row cannot be scored until the sheet is full — take all twelve counts, then split them into thirds. Its catalogue is far larger and its search matches loosely, so that number is only a rank.

One signal stays outside the score: sale badges. TeePublic sets the retail prices and runs sitewide discounts for much of the year, so badge share tracks the promo calendar rather than the phrase. Badge-heavy results mean buyers have learned to wait, and the artist royalty per unit is lower on a discounted sale — check the current royalty table in TeePublic's artist help pages, last checked September 2026.

Read the First Twenty Results for What Is Missing

Counting says how crowded a phrase is; opening listings says what the crowd is. Do this only for the two or three phrases you will draw — thirty minutes each. Open the first twenty of the same 48 tiles and record title, phrase matched, default product type and badge. Only two reach the sheet: the product type most of the twenty default to, and the phrase variant their titles match. Badge you read for the promo pattern and discard, since it gets no column.

If eighteen of twenty titles say funny nurse and none say night shift nurse, the smaller phrase is your entry point precisely because it is smaller. Where all twenty are typographic, an illustrated design wins attention before a word is read. Low competition never means nobody has uploaded a matching design; it means the phrase is thin in titles and its tag page has room.

Score Demand Against Competition

Be honest about the cost. Four signals run about ten minutes a phrase once the tag page means opening six products, so forty phrases is seven hours — nobody finishes that. Score demand for all forty, then count competition on the top twelve: two hours, or three end to end.

Demand starts from the dropdown behaviour in Method 1, then moves with the Planner range, so the score reflects buyers and not only sellers. Paste all forty into Keyword Planner's "Get search volume and forecasts" — the "Discover new keywords" path caps seeds and will not take the list. A new Google Ads account returns wide ranges, so read them as present-or-absent.

That adjustment is one point either way, floored at 1 and capped at 5. It stays modest because a Planner range measures the open web, not TeePublic's box: at best it lifts a never-suggested phrase to demand 2, which nets 1 even against the emptiest phrase.

Pick the Main Tag With the 0-3 Relevance Rubric

Here is where sellers stall. Net score is the wrong input for the main tag: a phrase can net 2 by being tiny and empty, and an empty tag page has no browsers. Use three gates: the candidate scores 3 on the rubric below; its tag page holds more than one screen; and of the phrases clearing both, it carries the highest demand.

Score candidates against the artwork, not the niche: shoppers arriving through a tag that misses it leave at once. A title is where a long phrase earns its match; a main tag is where you want a busy room. Both must score 3, so a score-2 audience phrase never takes the main slot.

Row zero is where accounts are lost. TeePublic removes designs referencing registered marks in titles or tags, and repeated removals escalate to account closure. Run any phrase naming a brand, show, team or character through USPTO Trademark Search at tmsearch.uspto.gov, successor to TESS, or follow the trademark checker guide. Not legal advice; last checked September 2026.

Tags can be edited later, so the main tag is not one-way — though TeePublic publishes nothing about how an edited listing is treated. Test on one low-value listing.

Build the Listing From the Winners

Relevance is the scarce thing, not slots. TeePublic's own rule is blunt: repeating the same word hurts your search results, so no two tags should be near-spellings of each other. "nurse humor" and "funny nurse" are one search dressed twice, and carrying both is worse than carrying one. Title and main tag are a different matter — separate fields doing separate jobs, so repeating a phrase across them costs nothing and is usually right. Tag mechanics sit in the tags guide.

That example spends ten distinct phrases, comfortably inside TeePublic's 6-to-15 guidance and with room to add more only if genuinely relevant ones remain. If they do not, stop: fifteen tags with four weak ones ranks worse than ten strong ones. Tag pages are indexed by Google, so a readable title brings buyers who never touch the search box.

Write a Separate Tag Set Per Product Type

Product type changes phrasing, which is what the default-product-type column is for.

• Tees and hoodies — bought as identity: occupation, role, life stage, in-group joke.
• Stickers — bought as objects on impulse, so nouns naming the thing depicted beat identity phrases; more in the sticker research guide.
• Mugs, wall art and phone cases — bought as gifts and decor, so occasion and room words earn slots a tee joke wastes.

Before committing a week of art, apply the size test from Method 1; shortlists sit in the niches roundup. Google Trends answers seasonality free — a phrase peaking in early May wants uploading in March, because approval and first sales eat time.

Worked Example: Night Shift Nurse

Harvest and score demand. Seed nurse logged out, walk a to z, reverse the strong ones, then add gift, retired and night shift. Night shift nurse surfaces under two trigger letters — base 3 — and Keyword Planner returns a non-zero range, so demand is 4.

Score competition on the top twelve. The readings below are illustrative: they show how a completed row converts to a score. Take your own on the day.

• Result depth about 104 designs, 26 rows at four across — Open
• Exact-phrase titles 4 of 48 — Open
• Redbubble printed count in the lowest third of the sheet — Open
• Tag page, reached by clicking the tag on a product, runs a few screens — Contested

Three Open of four, so competition is 5 minus 3, which is 2. Demand 4 minus competition 2 nets 2: design it. Run the same four on the bare seed nurse and you normally find the opposite — still loading past 500, an endless tag page, a Redbubble count in the top third: competition 5, which nets nothing.

Run the gate check before accepting the main tag. Night shift nurse scores 3 on relevance, its tag page runs a few screens, and it carries the highest demand of the phrases clearing both, so it takes the main slot. Had that page read Missing, the main tag would move to nurse coffee, the only other phrase scoring 3 here — and if that one is Missing too, ship your best score-3 phrase and expect no browse traffic.

Clear the title, then ship. The title goes through USPTO Trademark Search as well as the main tag, because a removal triggered by a title costs the same account as one triggered by a tag. Night Shift Nurse Coffee Then Chaos names no brand, show or character, which is what lets it through.

Read it back after ninety days. Views with no sales means the tags pull the wrong shoppers: swap score-1 tags for score-2 occasion phrases and leave the artwork alone. No views at all usually means an empty main tag page, so move to the next phrase clearing the gates. Change one variable per design, never re-tag something selling, and past fifty listings see the bulk upload comparison.

Tools and What Each Is Actually For

No single tool spans demand, competition and trademark safety.

Paid tooling earns its place when the trademark pass across forty titles is the step you keep skipping, or when one tag template decides whether you ship. Below a handful of designs a month, do both by hand. For Merch by Amazon an Amazon-catalogue tool reaches BSR data no TeePublic-side tool sees; for raw demand Keyword Planner beats any print-on-demand product.

FAQ