T-shirt Market Research: Finding Profitable Room Before Designing

00 · Thesis

“Untapped” is the wrong target. “Profitable room” is the target.

Empty market

No external demand plus no marketplace proof usually means there may be no buyer.

Huge clone market

A crowded first page of exact-match designs and low prices usually means the obvious version is saturated.

Profitable room

Visible demand plus repeated weak designs, a specific buyer, a better angle, and viable economics is worth testing.

IP kill switch

High trademark, copyright, publicity, brand, lyric, team, celebrity, character, or event-reference risk means skip or reframe.

Working definition: a researchable niche is “X person buying Y shirt for Z reason,” not “dogs,” “teachers,” “pickleball,” or any other broad topic.

01 · Process

A repeatable niche-discovery workflow.

Pick seed buyers the operator can understand or reach.

Start with local groups, recurring gift contexts, hobbies with insider vocabulary, professional identities, or pet/breed identities. Avoid broad supplier-blog categories unless the next move is narrowing.

Convert seeds into buyer-intent phrase clusters.

Generate 10–20 phrases such as [audience] gift shirt, [event] family shirts, [pet breed] mom shirt, "[exact phrase]" t-shirt, plus modifiers like subtle, personalized, not cheesy, team, review, retro, or minimalist.

Gather evidence by surface.

Separate marketplace purchase signals, search/trend language, community/review language, economics, and IP checks. Record query, source, date, URL or allowed screenshot, fact, and interpretation.

Inspect competitors for room, not just count.

For the top 50–100 relevant results, look for reviews, recency, price discipline, personalization, shipping deadlines, mockup quality, visual repetition, and expansion potential.

State the wedge before design.

Require one sentence: “This shirt is for [buyer] who wants [tone/identity/occasion], but current options are [failure], so this design will [specific improvement].”

Score, then choose the next action.

End each candidate with a decision: kill, watch, test, or design. The output of research is a decision, not an interesting idea list.

02 · Scorecard

Score demand, competition-room, and monetizability separately.

Demand

0.40 weight
Buyer intent, marketplace proof, seasonality, recurring need.

Room

0.30 weight
Weak incumbents, poor personalization, stale tropes, clear wedge.

Margin

0.30 weight
Channel economics, pricing power, bundles, local/group orders.

Decision bands

4.0–5.0

Design/test now, assuming IP risk is low.

3.0–3.9

Test 1–3 concepts or sharpen the angle first.

2.0–2.9

Watch, narrow, or gather more evidence.

< 2.0

Skip.

IP risk is not averaged in.

Evaluate trademark/IP/publicity risk separately as low, medium, or high. A high-risk phrase or concept should be skipped or reframed even if demand looks attractive.

Existing infringing listings are not permission. Use USPTO search and official platform policies before design work.

Buyer

A specific buyer and purchase moment are clear.

Demand

At least two independent surfaces show directional demand, and one is close to purchase behavior.

Room

Competitor inspection shows exploitable gaps rather than only raw result counts.

Economics

The channel supports contribution margin after production, platform/payment fees, shipping, discounts, ads, returns, and design time.

Risk

IP, quality, and ethical/community-fit risks are low or manageable.

03 · Patterns

Promising directions are narrow; avoid directions are usually obvious and crowded.

Promising if verified

Role + exact occasion: first-day teacher teams, graduation-year family shirts, coach gifts.
Hobby + demographic + insider joke: narrower than the generic hobby query, with repeatable variants.
Local/group-order niches: schools, fundraisers, neighborhoods, small businesses, clubs.
Personalization-heavy events: family trips, reunions, bachelorette groups, team parents.
Aesthetic remix with clear buyer language: a real visual wedge beyond “same phrase, different font.”

Avoid or reframe

Broad commodity queries like “funny teacher shirt,” “dog mom shirt,” “pickleball shirt,” or “vintage graphic tee.”
Pure trend or meme shirts with short windows and copycat rushes.
Fandom, sports-team, brand, lyric, celebrity, character, or protected event references.
Low-price clone markets that cannot support POD contribution margin.
Sensitive community topics where Clay cannot add informed value.

04 · Validation

Use cheap tests before polished design effort.

1. Paper pass

Score 5–10 candidates in the same worksheet so they can be compared honestly.

2. Rough mockups

Create fast concepts only for candidates above threshold. Do not polish weak hypotheses.

3. Intent test

Use a small marketplace/listing, landing page, organic post, Pinterest pin, paid click test, signup, preorder, or direct group-order ask.

Track: impressions, clicks, favorites/saves, add-to-carts, conversion, questions, refunds/quality issues, margin, and review language. Feed results back into the scorecard.

05 · Worksheet

A candidate row should make the next decision obvious.

Candidate niche/design direction:
Buyer:
Purchase moment: self / gift / event / group / identity / profession / other
Primary channel: Amazon / Etsy / Redbubble-TeePublic / Shopify-social / local-B2B
Primary keywords:
Long-tail variants:
Seasonality:
Marketplaces checked:

Evidence log
- Marketplace buyer proof:
- Search/trend proof:
- Community/review language:
- Competitor price band:
- Top listing sameness:
- Differentiation gap:
- Margin estimate:
- IP/legal risk:
- Quality/fulfillment risk:
- Ethics/community fit:

Scores
- Demand, 1–5:
- Competition-room, 1–5:
- Monetizability, 1–5:
- Weighted opportunity score:
- IP risk: low / medium / high
- Decision: kill / watch / test / design

Why this wins sentence:
“This shirt is for [buyer] who wants [tone/identity/occasion], but current options are [failure], so this design will [specific improvement].”

Next action:

06 · Caveats

What is settled enough, and what must be refreshed per candidate.

Settled enough for the process

Underserved sub-angle beats “zero competition.”
Marketplace, search, social, and community metrics are directional and must be triangulated.
First-page quality and exact-match density are better saturation checks than raw result counts alone.
IP checks belong before design work.
Research should end in kill, watch, test, or design.

Must be refreshed live

Marketplace counts, ranks, review recency, visible BSR/review data, and prices.
Exact platform fee, royalty, and policy wording.
Whether the operator can reach a specific buyer group cheaply.
Actual conversion rates for the operator's listings, pricing, style, and fulfillment quality.
Whether a phrase is trademark-safe in the specific apparel class and context. This is not legal advice.

Public/private posture: publish the methodology, scorecard, caveats, broad examples, and source list. Keep Clay's actual candidate niche list, private community details, live scraped observations, raw prompts, account output, and exact unverified market claims private.

07 · Sources

Source map carried forward from the research lanes.

These links support the workflow and evidence hierarchy. Marketplace and platform pages can be dynamic or gated; refresh exact fee, policy, ranking, and pricing claims before making business decisions.