1,117 responses, 5 models, one vertical: the first AI-visibility snapshot of Turkey's innerwear, hosiery and beachwear market

Vertical: Innerwear, hosiery & beachwear — Turkey
Method: A single point-in-time study of 1,117 responses — five large language models (GPT-4o-mini, Gemini, Claude, Perplexity, Grok), each asked 45 Turkish-language questions five times
Collection dates: 19–20 June 2026

Key terms: AI visibility, AI Visibility Score, innerwear brands, hosiery brands, beachwear brands, Turkish (yerli) underwear/swimwear brands, LLM brand recommendations, ChatGPT/Gemini/Claude/Perplexity/Grok brand recommendation, Generative Engine Optimization (GEO).

This report reproduces the real Turkish-language questions consumers ask AI assistants about underwear, socks, tights and swimwear, and measures how often, in what order, and across how many models five large language models name each brand. Because the same questions can be read with or without an origin restriction, the brands are ranked in three cuts — main, local and open — defined in Chapter 2. It is a single point-in-time study and should be read within the limitations in Chapter 10.

How to read this report — visibility is not quality

This report measures one thing: how often, how prominently, and across how many AI assistants a brand is named when people ask about innerwear, hosiery and beachwear in Turkey. A high score means these systems currently have a lot of information available about a brand and surface it readily — not that the brand is better, more comfortable, cheaper or higher quality. A low score, or a zero, means a brand is currently less visible to these systems — not that it is inferior. AI visibility reflects information availability and discovery, not product quality, fit, price or trustworthiness. Readers and shoppers should not treat an AI recommendation — or this report — as a verdict on whether a brand is “good” or “bad.” This study does not measure the accuracy, quality, or sentiment of any recommendation.

1. Executive Summary

One-sentence takeaway: AI visibility in Turkey’s innerwear, hosiery and beachwear market is led decisively by Penti and Ayyıldız, sits on a Turkish-majority but genuinely contested shelf, and is split cleanly by subcategory — a different brand leads underwear, socks and swimwear.

• Penti leads; Ayyıldız is a clear second. Penti tops the main leaderboard at 94.91, appearing in 41.3% of all responses across all five models, with an 8.35-point gap to Ayyıldız (86.56). Trendyol (75.9), Dagi (68.92) and Dündar Çorap (68.67) complete the top five.
• 28 brands qualified for the main ranking. Of 189 tracked brands, 28 cleared the 5% mention threshold across all 45 questions — 19 Turkish and 9 foreign. The two foreign names inside the top ten are Marks & Spencer (#7) and OYE Swimwear (#10). The average score of the 28 qualified brands is 56.1.
• The market is Turkish-led but not Turkish-only. On the 22 origin-neutral (open) questions, mentions split 59.3% Turkish / 40.7% foreign — the most balanced origin split Herm.io has recorded in any vertical to date. Foreign visibility here is real and broad (12 foreign brands qualify in the open cut), not confined to a single niche.
• The local shelf is overwhelmingly Turkish. When questions explicitly ask for Turkish/yerli brands (the 23 local questions), only 4 of 36 qualified brands are foreign (OYE Swimwear, Sand&Blue, Moeva London, Marks & Spencer). The assistants clearly recognise “yerli” as a constraint and answer it with domestic names.
• Subcategory leadership is split three ways. Penti leads innerwear (52.6% within the subcategory), Dündar Çorap leads hosiery (37.7%), and Ayyıldız leads swimwear/beachwear (58.5%). No single brand owns all three shelves.
• The models see the same market at very different breadths. Grok lists ~9.4 distinct brands per answer; GPT-4o-mini (3.43) and Claude Haiku (3.64) are the most selective. Despite that, Penti and Ayyıldız lead across models.
• Search was near-universal. 98.7% of all answers were web-grounded; four of five models searched on ~100% of answers and only Gemini fell below (93.1%). With almost no “memory-only” answers, this edition does not include a web-search-vs-memory comparison; see Chapter 5.
• Sources are brand-owned first. 45.3% of cited sources are brands’ own websites; the single most-cited domain is the marketplace trendyol.com (40.4% of responses), followed by brand sites ayyildiz.com.tr and penti.com.

Why it matters. Shoppers are shifting discovery from search engines and marketplaces to AI assistants. Whether an underwear, sock or swimwear brand appears in those assistants’ answers is becoming a discovery channel in its own right. This report captures the first snapshot (baseline) of this new “visibility shelf” for Turkey’s innerwear, hosiery and beachwear sector. A reminder: the numbers below describe visibility and information availability — not which brands are best.

2. Methodology

One-sentence takeaway: Five large language models were asked 45 Turkish questions five times each — with no system prompt — to produce 1,117 usable responses; brands were identified by automated extraction plus human validation and ranked with a three-component (45/30/25) score, computed separately for the main, local and open cuts.

2.1 Scope and models

The study queried five large language models directly via API with identical questions. All models received only the user question — no system prompt was used, temperature was not set manually (provider default), and no region/locale parameter was defined. Because the questions are in Turkish, models infer the Turkish market from language alone.

Scale: 45 questions × 5 repeats × 5 models = 1,125 responses; 8 were excluded (7 empty Gemini generations and 1 stuck Perplexity row), leaving 1,117 usable. Single run, 19–20 June 2026. The 45 questions are evenly split across three subcategories — innerwear (İç giyim), hosiery & socks (Çorap) and swimwear & beachwear (Mayo/Bikini), 15 each — and across three behavioural types — discovery, attribute and use_case, 15 each. Final brand dictionary: 189 brands (130 Turkish, 59 foreign).

The three market cuts. Rather than two side-by-side markets, this vertical is read in three cuts of the same questions:

2.2 Web-search configuration

Web search was offered as an enabled tool to all five models; each decided for itself whether to use it (Perplexity’s Sonar is search-grounded by design). Whether a response drew on web search or the model’s own knowledge was inferred by whether the provider returned any citation/source: a citation means “web-search,” none means “own knowledge.” This is a reasonable proxy, not direct proof of the model’s internal process. In this run 98.7% of answers were grounded (see Chapter 5).

2.3 Questions

The 45 questions were generated in a two-step process and then balanced across the three subcategories and three behavioural categories:

1. Seeding. We first asked LLMs to generate the questions and keywords people in Turkey might plausibly ask an assistant about buying underwear, socks/tights and swimwear.
2. Validation against real demand. We took those seeds to Google, YouTube and Ekşi Sözlük to find related real-world queries, and used Google’s “People also ask,” alsoasked.com and answerthepublic.com for additional question and keyword variants. We then merged the lists and selected the top 15 per subcategory by occurrence.

Of the 45, 23 are “local” (they explicitly ask for Turkish / yerli brands) and 22 are “open” (no origin restriction). The Turkish-vs-foreign comparison in Chapter 7 rests solely on the 22 open questions — the only fair basis for comparing Turkish and foreign brands.

2.4 Brand extraction and validation

Brands were extracted in two stages. (1) Automated extraction: emphasised brand names and citation domains were collected, and spelling/format variants merged into a single canonical name. (2) Human validation: the candidate list was reviewed by hand — real brands confirmed, non-brands (category words, generic phrases, fabric types, places) filtered out, aliases merged, and each brand labelled Turkish/foreign. Final dictionary: 189 brands. Per study scope, retailers and marketplaces (Trendyol, Hepsiburada) are kept as entities so the shelf is described as shoppers actually meet it; a brand-only view is given in §3.2.

Origin rule. A brand is Turkish if founded and headquartered in Turkey; foreign otherwise, even when widely sold in Turkey (e.g. Marks & Spencer, Calzedonia, Intimissimi, Victoria’s Secret).

Definitions.

• Mention: a validated brand (or alias) appears in the answer text — counted once per response (presence, not frequency), matched word-boundary aware, case- and diacritic-insensitive, longest-alias-first so a shorter name never double-counts inside a longer one.
• Position (MRR): mean reciprocal of the brand’s first-appearance rank within each answer (first = 1.0; second = 0.5; …), averaged only over the responses where the brand appears.
• Breadth: number of the 5 models (0–5) that mention the brand within that market.

2.5 AI Visibility Score (0–100)

Score = 0.45 × Mention + 0.30 × Position + 0.25 × Breadth

Each component is scaled so the leading qualified brand in each cut = 100, then combined with the weights. Qualification (≥5% rule): only brands mentioned in at least 5% of that cut’s responses are ranked (28 main, 36 local, 26 open). This threshold limits the risk that a rare brand which happens to appear first in a few answers inflates the position component. Unfiltered metrics for all 189 brands appear in Chapter 11.

2.6 Neutrality and self-exclusion

This is a market-wide, neutral study with no focus brand. To prevent any conflict of interest, all citations to Herm.io’s own domains were excluded from the source data before analysis, so the company’s own content neither appears among the most-cited domains nor influences any figure.

3. Overall Visibility Leaderboards

One-sentence takeaway: Penti and Ayyıldız lead the market in every cut; the difference between cuts is what happens behind them — the local shelf is almost entirely Turkish, the open shelf is genuinely contested by foreign brands.

As a reminder, these rank visibility, not product quality, fit, price or reliability.

3.1 Main leaderboard — 28 qualified brands

Average score of the 28 qualified brands: 56.1.

Reading. Penti maxes out the mention and breadth components and tops the table at 94.91, appearing in 41.3% of answers. Ayyıldız (86.56) is second on the strength of the highest position score in the top tier (MRR 0.61). Then come Trendyol (75.9), Dagi (68.92) and Dündar Çorap (68.67) — a notable case: Dündar is named in only 12.5% of answers but, when named, almost always first (MRR 0.66, the highest position component of any brand), which is why it sits fifth despite thin mentions. 25 of the 28 qualified brands are seen by all five models; the top ten are foreign only twice — Marks & Spencer (#7) and OYE Swimwear (#10).

3.2 Brand-only view (marketplaces removed)

Two entities in the table above are marketplaces rather than brands — Trendyol and Hepsiburada. They are kept in the headline ranking because that is how shoppers meet the shelf, but removing them gives a pure brand-versus-brand picture:

With Trendyol and Hepsiburada set aside, the top of the market is Penti, Ayyıldız, Dagi, Dündar Çorap and Suwen — all Turkish — with Marks & Spencer and OYE Swimwear the highest foreign names. (DeFacto and Koton are treated as brands, not marketplaces, and remain in both views.)

3.3 Local leaderboard — 36 qualified brands (Turkish-only questions)

When the 23 questions that explicitly ask for Turkish/yerli brands are isolated, the shelf gets longer (36 qualify) and almost entirely domestic.

Average score of the 36 qualified local brands: 52.6. Only 4 of the 36 are foreign.

Reading. Penti (93.55) and Ayyıldız (88.36) again lead, but the local cut elevates specialist domestic names: Dündar Çorap rises to #3 (75.27), and small, position-strong brands such as Haşema (#6, mentioned in 5.6% of answers but with MRR 0.79) and Woolona (#7, 5.1%, MRR 0.78) clear the bar mainly on how early and how broadly they are named. Their ranks should be read cautiously (see Chapter 10). The only foreign brands to survive an explicit “yerli” filter are OYE Swimwear, Sand&Blue, Moeva London and Marks & Spencer — all in the lower half.

4. Differences Between Models

One-sentence takeaway: All five models broadly agree on the leaders, but they differ enormously in breadth — Grok names nearly 9.4 brands per answer, GPT-4o-mini and Claude Haiku closer to 3.5.

4.1 Per-model behaviour summary

Web search was near-universal: four of the five models searched on essentially every answer; Gemini on 93.1%. The clearest difference is breadth: Grok lists ~9.4 brands per answer and reaches the longest tail, while GPT-4o-mini (3.43) and Claude Haiku (3.64) name a much more selective set. Perplexity (5.35) and Gemini (6.25) sit in between. Note Grok’s most-named entities are the marketplaces (Trendyol, Hepsiburada) rather than a single brand — a different answering style from the others.

4.2 Each model’s most-named brands

The tables below give each model’s five most-mentioned entities across its ~225 responses (rate = % of that model’s answers naming the brand). Penti or Ayyıldız tops four of the five models; Grok is the exception, led by Trendyol.

The practical reading: which brands appear at the very top is fairly consistent, but how many brands an answer contains — and whether a marketplace or a brand is named first — depends heavily on the model. A long-tail brand’s chance of being named is far higher in Grok than in the more selective GPT-4o-mini or Claude Haiku.

5. Search vs. Memory: Why This Edition Omits the Split

One-sentence takeaway: Almost every answer in this run used web search, so there is no meaningful “memory-only” segment to compare against — itself a finding.

Some earlier Herm.io editions split answers into web-grounded and memory-only segments that produced different leaders. That split does not exist in this run. About 98.7% of the 1,117 answers returned at least one citation or search result; the “own-knowledge” segment is a handful of responses — far too small to support a comparison. Four of the five models (GPT-4o-mini, Claude Haiku, Perplexity, Grok) searched on ~100% of answers, and only Gemini fell below 100%, at 93.1%.

Because almost all queries used web search, a separate web-search-vs-memory chapter would compare ~1,102 grounded answers against ~15 ungrounded ones — not a meaningful contrast — so it is intentionally omitted this edition. The takeaway is that, for Turkish innerwear, hosiery and beachwear questions, these assistants reach for live web content almost every time. That makes a current, well-structured web presence especially closely tied to visibility here (see Chapter 8). We will keep measuring the grounding rate each edition; if a memory-only segment grows, the comparison will return.

6. Category Ownership

One-sentence takeaway: No brand owns the whole vertical — Penti leads innerwear, Dündar Çorap leads hosiery, and Ayyıldız leads swimwear/beachwear — and question type reshuffles the order, with the marketplaces surfacing more on attribute and use-case questions.

6.1 By subcategory

Each subcategory is built from 15 questions. The tables show the most visible brands within each one (rate = relative to that subcategory’s responses).

Reading. The three shelves have three different leaders. Penti dominates innerwear (52.6%), well ahead of Dagi (41.6%) and Ayyıldız (34.9%). Dündar Çorap leads hosiery (37.7%), narrowly ahead of the marketplace Trendyol (36.4%) and Penti (28.3%) — the one subcategory a specialist owns outright. Ayyıldız leads swimwear/beachwear (58.5%), ahead of Penti (42.9%); this is also where foreign brands are most visible (Marks & Spencer, Calzedonia, OYE Swimwear, Sand&Blue all rank in the subcategory’s top ten).

6.2 By question type

The same questions, grouped by behavioural type — discovery (general “best/recommended”), attribute (specific qualities — seamless, bamboo, supportive, plus-size) and use_case (need-driven — postpartum, sport, sensitive skin).

Reading. Penti leads discovery (47.4%) and use-case (40.1%) questions, while the marketplace Trendyol leads attribute questions (40.1%) — when shoppers ask for a specific property, the assistants lean on the marketplace’s breadth of listings. Hepsiburada also rises on attribute and use-case questions, the same pattern: marketplaces gain ground when the question is specific rather than a general “recommend a brand.”

7. Open Market: Turkish vs. Foreign

One-sentence takeaway: Across origin-neutral questions, 59.3% of mentions go to Turkish brands and 40.7% to foreign — the most balanced split Herm.io has measured, with foreign brands competitive across the board rather than parked in one niche.

This chapter relies only on the 22 open questions with no origin restriction (546 responses) — the only fair basis for comparing Turkish and foreign brands.

7.1 Origin share

The Turkish lead is real but modest: 1,742 Turkish mentions to 1,195 foreign. Breadth tells the same story — 78 distinct Turkish brands surface in the open market versus 55 foreign. Compared with verticals where domestic brands dominate origin-neutral questions outright, this market is genuinely contested: foreign innerwear and swimwear names are recalled readily even when the question does not invite them.

7.2 Open-market top 15 brands (any origin)

7.3 The pattern: Turkish leads, foreign is never far behind

The individual ranking shows the structure the aggregate hides. The open-market top three are Turkish — Penti, Ayyıldız, Trendyol — but Marks & Spencer ranks #4, ahead of Dagi, and foreign names (Intimissimi, Vilebrequin, Adidas, Calzedonia, Victoria’s Secret) populate the rest of the top fifteen. So when a shopper asks an origin-neutral question, the answer leads with Turkish brands but routinely includes foreign ones — especially in lingerie (Intimissimi, Victoria’s Secret, Chantelle) and premium swimwear (Vilebrequin, OYE Swimwear). Whether that balance shifts is exactly the kind of movement a quarterly baseline is built to track.

8. The Discovery Ecosystem: Where AI Learns About Brands

One-sentence takeaway: AI’s picture of this market rests first on brands’ own websites (45.3% of cited sources) and a large marketplace presence — led by trendyol.com — rather than on editorial or review content.

When an assistant answers with web search, it draws on the content available to it at that moment. We collected the source addresses returned by the models, reduced them to domains, and ranked them by how many distinct answers cited each — a map of where these systems read about innerwear, hosiery and beachwear brands.

8.1 The 9 most-cited domains

The single most-cited domain is the marketplace trendyol.com (cited by 40.4% of responses), but it is followed immediately by a wall of brand-owned sites — ayyildiz.com.tr, penti.com, marksandspencer.com.tr, dagi.com.tr, suwen.com.tr, dundarcorap.com — interleaved with a second marketplace (hepsiburada.com), shopping/editorial sites (glingerie.com.tr, oggusto.com, onedio.com) and forums (eksisozluk.com, reddit.com).

8.2 Source-type mix

8.3 The brand-owned finding

The most striking pattern is concentration on official and marketplace sites: 45.3% of citations point to brands’ own domains and another 12.1% to retailers/marketplaces, with a large 37.0% “other” bucket (brand and independent sites not in the dictionary) and only small shares for social/UGC (3.7%) and editorial/news/blog (1.8%). The practical reading (a description, not advice): in this market, a current, machine-readable brand site — plus presence on the big marketplaces — appears closely associated with AI visibility, more so than editorial or review coverage.

9. What the Patterns Suggest

One-sentence takeaway: The most visible brands tend to share three things — coverage across all five models, a strong own-site and marketplace web presence, and clear leadership in at least one subcategory.

This chapter is a neutral reading of the data. It describes patterns associated with visibility; it is not advice, a service, or a recommendation, and visibility remains separate from quality.

1) Breadth is a baseline condition. 25 of the 28 qualified main-market brands appear in all five models. Brands seen by only three or four models sit lower largely for this reason — broad model coverage is what the most visible brands share.

2) Mention and position are different things. Some brands are named often but late (Trendyol, Hepsiburada — high mentions, low MRR); others are named rarely but early (Dündar Çorap, Haşema, Woolona — thin mentions, very high MRR). A brand can clear the 5% bar mainly on position and breadth; “named often” and “named early” are independent signals.

3) Web presence tracks with the dominant scenario. With 98.7% of answers grounded in web search, and brand-owned plus marketplace sites the largest source types, a current and structured web presence is the factor most closely associated with visibility here.

4) Subcategory is where leadership is actually won. No brand leads all three shelves. Penti owns innerwear, Dündar Çorap owns hosiery, Ayyıldız owns swimwear/beachwear. A brand’s context — which subcategory, which question type, which cut — matters as much as its overall rank.

5) Origin is contested, not settled. Unlike verticals where domestic brands dominate origin-neutral questions, foreign innerwear and swimwear brands are recalled readily in the open market (40.7% of mentions). The local/open gap shows the assistants can distinguish “yerli” requests — the open balance is therefore a real competitive signal, not an artefact.

How a brand can locate itself in the data. Using the dataset (available on request), a brand can read four things in order: whether it appears in all five models; how often it is mentioned within its subcategory and cut; how early it is named when mentioned; and which model or question type it is weakest in. These readings show where a visibility gap sits — without saying anything about whether its product is good.

10. Limitations and Notes

One-sentence takeaway: This is a single, point-in-time snapshot of one Turkish vertical; read the figures within the limitations below.

• Single, point-in-time measurement. Collected in one run on 19–20 June 2026 — a snapshot, not a trend. This is the first (baseline) edition; quarterly editions will enable trend analysis.
• Small-sample position inflation. A few brands clear the 5% bar but rank high mainly on position and breadth despite thin mentions (e.g. Dündar Çorap, Haşema, Woolona). Read their ranks cautiously.
• Small score gaps are not meaningful. In the packed middle of each leaderboard, differences of a point or two should not be over-interpreted; only larger gaps (such as Penti’s lead) are robust.
• Visibility is not quality. This report measures only mention and position. It does not assess whether a brand was described positively, whether a recommendation was accurate, or whether a product is good. A score is not an endorsement.
• Models are probabilistic. The same question can produce different answers; five repeats reduce but do not eliminate this.
• Web search depends on momentary conditions — the content indexed and served at measurement time. “Web search vs. own knowledge” is a proxy based on whether a citation accompanied the answer.
• Precision over recall. Brand matching was cautious to avoid false positives; brands mentioned indirectly may be undercounted. There is no sentiment layer — the study is strictly quantitative.
• Source-type mix is heuristic. Domain types were auto-classified (brand-owned detected by matching the domain to a brand); the large “other” bucket is brand/independent sites not in the dictionary, so treat the mix as indicative. Source coverage also varies by model.
• Origin classification involves judgment. A brand is Turkish only if founded and headquartered in Turkey; some widely-sold names are foreign-origin even with a large local presence.
• Retailers/marketplaces are included as entities per study scope (Trendyol, Hepsiburada); they compete differently from direct brands, so §3.2 gives a brand-only view. DeFacto and Koton are treated as brands.
• Output-length limits and model versions differ and date quickly; findings are specific to the versions in Chapter 2 as of June 2026.

11. Appendix & Data

11.1 All questions (45) and their categories

Distribution: 15 discovery, 15 attribute, 15 use_case; 23 local (Turkish-only) and 22 open (origin-unrestricted); 15 questions in each of innerwear, hosiery and beachwear.

11.2 All tracked brands (189)

11.3 Data availability

The study tracked 189 brands; of these, 28 cleared the 5% threshold in the main market (with 36 qualifying in the local cut and 26 in the open cut). To support scrutiny and reproduction, the underlying data is available on request: response-level brand mentions, all three qualified leaderboards, full unfiltered brand metrics, per-model and per-subcategory breakdowns, the open-market breakdown, the top source domains, and the methodology and data dictionary.

All figures reflect a single point-in-time run (19–20 June 2026) and are subject to the limitations in Chapter 10. This is the first (baseline) measurement; the study will be repeated quarterly.

11.4 About Herm.io & disclosure

Herm.io is a consumer-behavior and marketing-data company. We study how people discover and choose brands so that brands can reach the right customers. This report is part of our public research and is a recurring quarterly study.

Disclosure & neutrality. Herm.io does not sell SEO or GEO (search/AI-ranking) services, and this report does not recommend any. No brand paid to be included, ranked, or described, and a brand’s score is not an endorsement or a judgment of its quality — it is a measure of visibility only. To keep the analysis objective, all citations to Herm.io’s own domains were excluded from the source data. Brands that want to understand their position in the data are welcome to book a call for a neutral walkthrough; this is advisory and free.