Vertical: Evening & occasion wear (abiye / davet giyim), Turkey
Method: A single point-in-time study of 999 responses — five large language models (Claude, Gemini, ChatGPT, Perplexity, Grok), each asked 40 Turkish-language questions up to five times
Collection date: 2 October 2026
Edition: 2 of a quarterly series. Edition 1 (July 2026) asked the same 40 questions; two of its five models were replaced for this edition.
Key terms: AI visibility, AI Visibility Score, evening wear brands, occasion wear brands, abiye / davet elbisesi, Turkish (yerli) evening-wear 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 evening dresses, abiye and occasion wear, and measures how often, in what order, and across how many models five large language models name each brand, retailer and marketplace. Entities are ranked separately within six cuts of the same questions — all questions, origin-constrained, origin-neutral, and three question types — defined in Chapter 2. It is a single point-in-time study and should be read within the limitations in Chapter 11.
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 evening and occasion wear in Turkey. A high score means these systems currently have a lot of information available about a brand and surface it readily. It does not mean the brand is better, more elegant, 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: Trendyol leads the general board and Carmen is the most visible entity that is not a marketplace, on a shelf that is entirely domestic when the question asks for Turkish brands and mostly domestic when it does not — domestic entities take 100.00% of qualifying mentions on the 20 origin-constrained questions and 83.21% on the 20 origin-neutral ones (foreign 0.00% and 16.79%), with origin on file for 2,485 of 2,485 and 1,769 of 1,769.
- Trendyol leads; Carmen is second. Trendyol, a marketplace, tops the general board at 100.00, named in 42.44% of all 999 responses (424) and by all five models; it holds the cut’s highest mention rate and its highest MRR, 0.4766. Carmen (93.77) is named in 38.44% (384 of 999). The ZKS (79.74), ADL (73.23) and Abiyefon (69.13) complete the top five. ADL is named slightly more often than The ZKS, in 26.03% of responses against 25.83%, but later: its MRR is 0.3279 against The ZKS’s 0.4347.
- 38 of 393 tracked entities qualify on the general cut — 35 domestic and 3 foreign — with an average score of 52.64. The foreign qualifiers are Mango (#14, 53.01), Zara (#22, 47.58) and Oleg Cassini (#28, 42.29). Three marketplaces — Trendyol (#1), Modanisa (#11) and Hepsiburada (#13) — rank alongside the brands, because they are what the assistants name. With the three removed, 35 entities remain and Carmen heads the brand-only board.
- The question changes how domestic the shelf is, not who leads it. On the 20 origin-constrained questions (499 responses), domestic entities take 100.00% of qualifying mentions and foreign entities 0.00%, with origin on file for 2,485 of 2,485. On the 20 origin-neutral questions (500 responses), domestic entities take 83.21% and foreign entities 16.79%, with origin on file for 1,769 of 1,769; 5 of the 30 entities in that share are foreign. Origin is assigned by brand home this edition and by a different rule in July, so no origin share here is compared with July’s.
- The assistants agree on Carmen, not on how long an answer is. Carmen is in every model’s top three, and it is the most-named entity for Gemini 3.6 Flash (47.50%) and Perplexity Sonar (37.50%). Trendyol is the most-named for ChatGPT 5.6 Luna Pro (35.68%) and Grok 4.3 (87.50%), and The ZKS for Claude Haiku 4.5 (27.00%). Gemini 3.6 Flash names 12.38 entities per answer and Grok 4.3 11.40, against 4.44 for Claude Haiku 4.5.
- Eleven qualifying ranks lean on one model. Each draws at least 60% of its general-cut mentions from a single model. Gemini 3.6 Flash, which joined the panel this edition, supplies 85 of İpekyol’s 109 mentions and 57 of Oleg Cassini’s 72. ChatGPT 5.6 Luna Pro, also new, supplies all 50 of Berneva’s. Grok 4.3 supplies 111 of Hepsiburada’s 178.
- Grounding was near-complete for four models and partial for one. 83.58% of all 999 responses were web-grounded. Claude Haiku 4.5, Perplexity Sonar and Grok 4.3 grounded 100.00% of their 200 answers and ChatGPT 5.6 Luna Pro 97.99% of its 199. Gemini 3.6 Flash grounded 20.00%: 40 of its answers report a search, and the other 160 returned no grounding metadata at all — which is not evidence that it answered from memory.
- Against July, The ZKS rose most and Iconas Line fell most — on a panel that changed. On the same 40 questions, with July’s answers counted the way this edition counts and with Gemini 2.5 Flash-Lite and GPT-4o-mini replaced by Gemini 3.6 Flash and ChatGPT 5.6 Luna Pro, The ZKS moved from 6.82% of 997 responses (68) to 25.83% of 999 (258), +19.01 points, and Iconas Line from 14.64% (146) to 6.11% (61), -8.54. The ZKS is in the top ten of all three models that answered in both editions (Chapter 8).
- trendyol.com is the most widely cited domain. It is cited in 84 of 200 question-model cells (42.00%), ahead of carmen.com.tr (69 cells, 34.50%) and thezks.com (59, 29.50%). Claude Haiku 4.5 returned exactly 5 citations on every one of its answers, so its domain counts are capped and do not compare with the other models’.
Why it matters. Shoppers increasingly ask AI assistants which evening dress to buy and where to buy it, and whether a brand appears in those answers is becoming a discovery channel in its own right. July’s edition set the baseline; this one repeats its 40 questions word for word, which makes a first reading of movement possible — with the confounds stated beside it, because two of the five assistants answering are not the ones that answered in July. The clearest finding is a structure rather than a movement: a marketplace leads, a group of domestic occasion-wear specialists sits behind it, and the shelf stays mostly domestic whether or not the question asks for Turkish brands. A reminder: the numbers below describe visibility and information availability — not which brands are best.
2. Methodology
One-sentence takeaway: Five models answered 40 Turkish questions up to five times each, giving 999 responses across 200 question-model cells; 393 entities were scored within each cut, and since July two models, the counting rule, the origin rule and the brand-only view have changed.
2.1 Scope and models
The study queried five large language models through their APIs with identical Turkish-language questions. Each of the 40 questions was put to each model up to five times, giving 200 question-model cells and 999 responses analyzed, all collected on 2 October 2026. Four models returned 200 responses each; ChatGPT 5.6 Luna Pro returned 199. The questions are in Turkish and models infer the market from them. The deployment’s system prompt and locale parameter were not recorded with this run, so this report makes no statement about either.
| Model | API name | Responses | In July's panel |
|---|---|---|---|
| Claude Haiku 4.5 | anthropic/claude-haiku-4.5 | 200 | Yes |
| Gemini 3.6 Flash | gemini-3.6-flash | 200 | No — joined this edition |
| ChatGPT 5.6 Luna Pro | openai/gpt-5.6-luna-pro | 199 | No — joined this edition |
| Perplexity Sonar | perplexity/sonar | 200 | Yes |
| Grok 4.3 | x-ai/grok-4.3 | 200 | Yes |
2.2 Web-search configuration
Every model was asked to search the web. Whether it did is a separate question, and the table below keeps the two apart: the first column is what the deployment asked of each model; the rest is what each model returned.
| Model | Asked to search | Answers | Web-grounded | Citations | Answer modes |
|---|---|---|---|---|---|
| Claude Haiku 4.5 | Yes | 200 | 100.00% | 1,000 | searched 200 |
| Gemini 3.6 Flash | Yes | 200 | 20.00% | 388 | searched 40, unknown 160 |
| ChatGPT 5.6 Luna Pro | Yes | 199 | 97.99% | 937 | searched 195, unknown 5 |
| Perplexity Sonar | Yes | 200 | 100.00% | 5,930 | searched 200 |
| Grok 4.3 | Yes | 200 | 100.00% | 1,051 | searched 200 |
Overall, 83.58% of the 999 responses were web-grounded. Claude Haiku 4.5, Perplexity Sonar and Grok 4.3 grounded 100.00% of their answers. ChatGPT 5.6 Luna Pro grounded 97.99%: 195 of its 199 analysed answers report a search and the other 4 carry no grounding metadata. Gemini 3.6 Flash grounded 20.00%: 40 answers report a search, and the other 160 carry no grounding metadata of any kind. An “unknown” answer is not a memory answer. The model was asked to search, and it is not known whether it did; the answer may rest on retrieval that was not reported, on the model’s own knowledge, or on both. This report never treats those answers as answers from memory.
July’s edition included a chapter on search versus memory, explaining why it did not split the two. This edition does not repeat it. Every model was asked to search, so there is no memory segment by design, and nearly all of the ungrounded answers are one model’s, of unknown provenance. Splitting the board by grounding would compare Gemini 3.6 Flash with the other four models, not search with memory.
A citation cap, not a measurement. Claude Haiku 4.5 returned exactly 5 citations on every one of its 200 answers. Its domain counts are truncated at that ceiling and do not compare with an uncapped model’s; Chapter 9 reads them that way.
2.3 Questions and cuts
All 40 questions are the ones asked in July, carried verbatim: none was added, none retired and none reworded. Each is typed by what it asks — discovery (a general “which brand?” or “where to buy?”), attribute (a specific property, such as a fabric, a plus-size fit or a dress that hides the tummy and hips) or use_case (an occasion or need, such as a henna night, a graduation or a wedding guest’s outfit) — and scoped by whether it constrains origin. A constrained question asks for Turkish brands (“iyi bir yerli abiye markası var mı” — is there a good Turkish evening-dress brand?); a neutral one does not (“en iyi abiye markaları” — the best evening-dress brands).
| Question type | Origin-neutral | Origin-constrained | Total |
|---|---|---|---|
| Discovery | 7 | 7 | 14 |
| Attribute | 7 | 6 | 13 |
| Use case | 6 | 7 | 13 |
Every metric is computed separately within each of six cuts:
| Cut | Questions | Models | Responses | Qualified | Distinct entities |
|---|---|---|---|---|---|
| All questions | 40 | 5 | 999 | 38 | 393 |
| Origin-constrained questions | 20 | 5 | 499 | 39 | 285 |
| Origin-neutral questions | 20 | 5 | 500 | 34 | 315 |
| Discovery questions | 14 | 5 | 350 | 53 | 291 |
| Attribute questions | 13 | 5 | 325 | 34 | 220 |
| Use-case questions | 13 | 5 | 324 | 32 | 226 |
July called the first three cuts Main, Local and Open. The all-questions cut is referred to in this report as the general cut.
July’s questions carried no topical subcategories, and July read the three question types instead; this edition reads them in Chapter 6, each with its own leaderboard. Under this series’ rules a subcategory is published as a cut only when the study declares it before collection. This study did not, so no subcategory figure appears in this report. Chapter 7 returns to this.
2.4 Entity resolution
What counts as a mention. A second language model reads each answer and lists each brand or company it names, once, in order of first appearance, up to 25 per answer. Product categories, ingredients and cities are excluded. A name that appears only inside a web address is not listed, and a shop named only as where to buy is often left out. Position (MRR) is the mean of 1 ÷ the entity’s first-appearance position, over the answers that name it.
Every listed name is then matched against Herm.io’s brand registry. A name that matches a registered brand or one of its aliases is counted directly and never becomes a candidate. Names that do not match are reviewed as candidates: 401 this period. Of those, 359 were resolved to a brand, 18 were given a verdict by a language model, and 24 were answered by a stored verdict from earlier work. These figures count the work, not the field, and do not reconcile to the 393 entities tracked.
Twenty-four candidates were rejected as not a brand. One of them was a single item naming more than one tracked brand; it was split, and each brand in it was credited. The rules behind the rest are easiest to see in examples.
A brand written only as a web address is not counted as that brand. “Carmen.com.tr” appears as a name in 11 responses. Written that way, it is not counted as Carmen. Most of the frequent rejections this edition are names of this kind — a shop or brand given as its web address rather than by name.
Some short names mean a specific brand in this study. Two short spellings are redirected before any verdict is asked for, because in this market they mean a different brand than the global registry would assume: “nedo” is counted as Nedret Taciroğlu, and “rue” as RUE Les Créateurs. The mentions move to those brands; the entity count does not change. No name is suppressed in this edition.
Resemblance is not identity. Among entities named in at least three answers, 12 near-name pairs were flagged — Trendyolmilla and Trendyol, Vakkorama and Vakko, and Beymen Club and Beymen among them. No pair shares a website on file, and each is tracked as two entities. All 12 are listed in Chapter 11.
The result is 393 entities: 135 brands, 193 brands that are also retailers, 37 retailers, 13 marketplaces, 6 publishers, 5 sub-brands and 2 licensed brands; 2 have no entity type on file. Origin is on file as domestic for 292 and foreign for 61; 40 have no origin on file. The two untyped entities, Didem Soydan and Armut.com, are also the two provisional ones: research never resolved them, and they may duplicate a brand under another spelling. No entity has a product category on file, so scope could not be resolved for any of them, and all 393 are eligible to rank.
2.5 AI Visibility Score (0–100)
Score = 0.45 × Mention + 0.3 × Position + 0.25 × Breadth
Mention is the share of the cut’s responses that name the entity. Position is the entity’s MRR, as defined above — first is 1, second 0.5, and so on. Breadth is the number of models, of five, that name it. Each component is scaled to the highest value among qualified entities in that cut and is never clipped. An entity qualifies for ranking on a cut if it is named in at least 5% of that cut’s responses; there is no absolute mention floor.
Worked example — Trendyol, general cut. Trendyol is named in 424 of 999 responses, a mention rate of 42.44%. That is the highest mention rate among qualified entities in the cut, so its mention component is 100.00. Its MRR of 0.4766 is also the cut maximum, so its position component is 100.00. It is named by 5 of 5 models, so breadth is 100.00. The composite is 0.45 × 100.00 + 0.3 × 100.00 + 0.25 × 100.00 = 100.00. The components are computed from unrounded inputs (a mention rate of 0.424424424 and an MRR of 0.476619658, each set against itself); the four-decimal MRR printed on this page is for reading, not for the arithmetic.
Trendyol scores 100.00 because it holds the cut’s highest value on all three components, so every other score on the general board is read against a marketplace. Because the maxima come from qualified entities only, an unqualified entity can exceed 100 on a component: 88 of 1,730 entity-cut rows do (5.10%), all of them unqualified — 5.90% of the 1,500 unqualified rows. A value above 100 says the entity beats the cut leader on that one component while being named too rarely to rank.
A score belongs to one cut’s scale. The scale moves whenever the leader moves, so an AI Visibility Score is never compared across cuts or verticals. Across editions it is compared in one place only: the general board in Chapter 3 shows each entity’s rank and score change against July’s answers re-scored the way this edition scores (section 2.6). Chapter 8 compares raw mention rate. The maxima behind every table are published in the report’s data file.
2.6 What changed since July, and what is comparable
| Edition 1 — July 2026 | Edition 2 — October 2026 | |
|---|---|---|
| Questions | 40 | The same 40, carried verbatim; none added or retired |
| Responses analyzed | 997 | 999 |
| Models carried | Claude Haiku 4.5, Perplexity Sonar, Grok 4.3 | Claude Haiku 4.5, Perplexity Sonar, Grok 4.3 |
| Models replaced | Gemini 2.5 Flash-Lite, GPT-4o-mini | Gemini 3.6 Flash, ChatGPT 5.6 Luna Pro |
| What counts as a mention | A validated brand name appearing in the answer text | A brand or company named in the answer, as listed by a second language model; not one that appears only in a web address |
| Origin rule | Founded and headquartered in Turkey | Brand home: the country the brand itself is operated from |
| Brand-only view removes | Marketplaces and retailers | Marketplaces only; retailers stay |
| Comparable across editions | — | Raw mention rate on the general cut, and the general board's rank and score change against July re-scored the same way — each with the panel change stated beside it. Not comparable: origin shares. |
Four changes matter for any comparison with July. The panel: three models are carried and two were replaced, and nothing in this report corrects for that. The counting rule: July’s page counted a validated brand name wherever it appeared in the answer text; this edition counts a brand or company only where a second language model lists it as named in the answer, and not where it appears only in a web address. The two count differently, and the brand registry has also changed since July, so a July level in this report can differ from July’s page for any entity. So that both sides of every comparison count a brand the same way, July’s levels in this report are July’s answers, counted the way this edition counts — not the figures July’s page printed. The origin rule: see 2.7. The brand-only view: July’s removed marketplaces and retailers; this edition’s removes marketplaces only, so retailers such as Sefamerve and Beymen stay in it (section 3.3).
Change on the leaderboard. The general board in Chapter 3 shows, beside each entity, how many places it moved and how much its score changed since July. Both are measured against July’s board rebuilt the way this edition builds its own — July’s answers and models, this edition’s brand registry and scoring — not against the figures July’s page printed, which counted differently. Besides the entity itself, two things move them, and nothing adjusts for either: the two models replaced between editions, and, for the score only, the leader. A score is a share of its cut’s leader, so when the leader is named more or less often every other score moves with it; rank change does not have that problem. 15 entities on this board were not on July’s, and 8 that were on July’s are not on this one.
2.7 Origin, neutrality and self-exclusion
Origin is the brand home: the country the brand itself is operated from, which is not always its owner’s. Fendi Casa is Italian though LVMH is French; Cetaphil is American though Galderma is Swiss. Where no separate brand home can be established, the parent’s headquarters stands in. Local-market naming variants are folded into their parent before origin is read, so a brand named after its market subsidiary is not recorded as domestic. Tables show ISO country codes; shares use a domestic/foreign rollup. Entities with no origin on file stay in every table, are excluded from every origin share, and are counted — never defaulted.
July used a different rule. July’s edition counted a brand as Turkish only if it was founded and headquartered in Turkey; this edition assigns every origin by brand home. An entity can therefore be coded differently in the two editions because the rule and the research behind it changed, not because the market did. Within this edition the rule is published here beside every origin figure, which is what makes a domestic-versus-foreign comparison safe; across the change it is not, and no origin share in this report is compared with July’s. July also published a Turkish-versus-foreign split on its origin-neutral questions, under its own rule and on its own basis. It is not set beside the shares in this report.
Marketplaces and retailers rank in every leaderboard, because they are what the assistants actually name. The brand-only view in section 3.3 removes the three qualifying marketplaces — Hepsiburada, Modanisa and Trendyol — by name and changes nothing else. Retailers, Sefamerve, Beymen, Vakko, LC Waikiki and DCEY among the qualifiers, are retained everywhere, as are brands that also run their own stores.
Origin shares are read on a brand-only, qualified basis. The headline share is computed over qualified entities with origin on file, with marketplaces and publishers left out — a marketplace is where a brand is bought and a publisher is where it is written about — and own-label retailers kept in. It is always printed with its coverage: the qualifying mentions with origin on file. The same arithmetic over every tracked entity, qualified or not, is published as a sanity check; it describes the long tail, not the shelf a reader sees. This basis applies to shares only; it does not change who ranks.
Neutrality and self-exclusion. This is a market-wide study with no focus brand. Herm.io’s own domain is excluded from the source data before analysis, so the company’s own content neither appears among the cited domains nor influences any figure.
3. Overall Visibility Leaderboard
One-sentence takeaway: Thirty-eight of 393 tracked entities qualify on the general cut; the marketplace Trendyol leads on frequency and position alike, Carmen is second at 93.77, and Mango, Zara and Oleg Cassini are the three foreign qualifiers.
The list below ranks the 38 entities named in at least 5% of all 999 responses, by AI Visibility Score. As a reminder, it ranks visibility, not quality.
Average score of the 38 qualified entities: 52.64. Thirty-five are domestic and three are foreign.
AI Visibility Score — qualified entities, all 40 questions
3.1 Tier narrative
The top two (100.00, 93.77). Trendyol is named in 42.44% of responses (424 of 999), by all five models, and earliest, with an MRR of 0.4766: it holds the cut’s highest value on every component. Carmen is named in 38.44% (384 of 999), also by all five models, with an MRR of 0.4452.
The ZKS to AlfaBeta (79.74–68.22). Position separates these four more than frequency does. ADL is named in 26.03% of responses (260 of 999) and The ZKS in 25.83% (258 of 999), but The ZKS ranks third on an MRR of 0.4347 against ADL’s 0.3279. AlfaBeta is named more often than Abiyefon, in 24.22% (242) against 23.32% (233), and ranks one place lower, with an MRR of 0.2786 against 0.3083.
Koton to Beymen (67.38–61.80). Koton, Sefamerve and Tarık Ediz follow. Tarık Ediz ranks ninth on 15.62% (156 of 999), carried by an MRR of 0.3747. Beymen, a retailer, ranks tenth on 7.61% (76 of 999) — by far the lowest mention rate in the top ten — because its MRR of 0.4565 is second only to Trendyol’s: when it is named, it is named early.
The middle of the board (60.45–46.13). Two marketplaces open this band. Modanisa (#11) and Hepsiburada (#13) are named often — 18.82% (188 of 999) and 17.82% (178 of 999) — but late, with MRRs of 0.2461 and 0.2467, and Hepsiburada by 4 of 5 models. Mango, Spanish, is 14th at 53.01, the highest-placed foreign entity, on 6.51% (65 of 999) with an MRR of 0.3354. Zara, also Spanish, is 22nd at 47.58 on 8.31% (83 of 999). Scores in this band sit close together, and differences of a point or less carry no meaning.
The tail of the board (44.04–20.92). Rengin to DCEY. Oleg Cassini, American, is 28th at 42.29 on 7.21% (72 of 999), named by 3 of 5 models. Breadth thins toward the bottom: Sagaza (33rd, 6.61%) is named by 2 models, Berneva (37th, 5.01%) by 1, and DCEY (38th, 5.31%) by 2, with the board’s lowest MRR, 0.0841. By Saygı and Berneva each qualify on 50 of 999 responses (5.01%), the minimum on this board. Chapter 5 shows which of these ranks rest mainly on one model. The remaining tracked entities sit below the threshold, and the appendix lists them all.
3.2 The same board by mention rate
Scores reward prominence as well as frequency, so the ranking by raw mention rate is a different list, and the difference is informative.
| # | Entity | Mention rate | Mentions | Score | Rank by score | Δ (+ = higher on score) |
|---|---|---|---|---|---|---|
| 1 | Trendyol | 42.44% | 424 | 100.00 | 1 | +0 |
| 2 | Carmen | 38.44% | 384 | 93.77 | 2 | +0 |
| 3 | ADL | 26.03% | 260 | 73.23 | 4 | -1 |
| 4 | The ZKS | 25.83% | 258 | 79.74 | 3 | +1 |
| 5 | AlfaBeta | 24.22% | 242 | 68.22 | 6 | -1 |
| 6 | Abiyefon | 23.32% | 233 | 69.13 | 5 | +1 |
| 7 | Sefamerve | 21.32% | 213 | 66.64 | 8 | -1 |
| 8 | Koton | 19.62% | 196 | 67.38 | 7 | +1 |
| 9 | Modanisa | 18.82% | 188 | 60.45 | 11 | -2 |
| 10 | Hepsiburada | 17.82% | 178 | 54.42 | 13 | -3 |
| 11 | Tarık Ediz | 15.62% | 156 | 65.14 | 9 | +2 |
| 12 | GIZIA | 13.91% | 139 | 56.61 | 12 | +0 |
Trendyol and Carmen are first and second either way. Below them, three pairs swap places — ADL and The ZKS, AlfaBeta and Abiyefon, Sefamerve and Koton — each time in favor of the entity named earlier. The marketplaces lose most to position: Modanisa is 9th by volume and 11th by score, Hepsiburada 10th and 13th. Tarık Ediz goes the other way, 11th by volume and 9th by score. GIZIA is 12th on both.
3.3 Brand-only view
Removing the three qualifying marketplaces — Hepsiburada, Modanisa and Trendyol, named explicitly, with nothing else changed — leaves 35 entities. Scores are unchanged, because the scale is still the general cut’s, which Trendyol set; entities ranked below a removed marketplace move up, and Carmen heads this board. Sefamerve, Beymen, Vakko, LC Waikiki and DCEY stay: they are retailers, not marketplaces. July’s brand-only view also removed retailers, so the two views are not the same list (section 2.6).
| # | Entity | Origin | Type | Score | Mention rate |
|---|---|---|---|---|---|
| 1 | Carmen | TR | Brand & retailer | 93.77 | 38.44% |
| 2 | The ZKS | TR | Brand | 79.74 | 25.83% |
| 3 | ADL | TR | Brand | 73.23 | 26.03% |
| 4 | Abiyefon | TR | Brand & retailer | 69.13 | 23.32% |
| 5 | AlfaBeta | TR | Brand & retailer | 68.22 | 24.22% |
| 6 | Koton | TR | Brand | 67.38 | 19.62% |
| 7 | Sefamerve | TR | Retailer | 66.64 | 21.32% |
| 8 | Tarık Ediz | TR | Brand & retailer | 65.14 | 15.62% |
| 9 | Beymen | TR | Retailer | 61.80 | 7.61% |
| 10 | GIZIA | TR | Brand | 56.61 | 13.91% |
| 11 | Mango | ES | Brand | 53.01 | 6.51% |
| 12 | Vakko | TR | Retailer | 52.86 | 10.71% |
| 13 | Roman | TR | Brand | 51.60 | 10.41% |
| 14 | Kenzel | TR | Brand | 51.40 | 12.01% |
| 15 | ModaSelvim | TR | Brand & retailer | 50.82 | 6.31% |
| 16 | Setre | TR | Brand & retailer | 50.10 | 9.61% |
| 17 | İpekyol | TR | Brand | 49.74 | 10.91% |
| 18 | EKOL | TR | Brand & retailer | 47.93 | 6.51% |
| 19 | Zara | ES | Brand | 47.58 | 8.31% |
| 20 | Lafaba | TR | Brand | 47.00 | 7.71% |
| 21 | Cengiz Aktürk | TR | Brand & retailer | 46.27 | 11.11% |
| 22 | Nocturne | TR | Brand | 46.13 | 10.91% |
| 23 | Rengin | TR | Brand | 44.04 | 9.41% |
| 24 | Armine | TR | Brand | 44.01 | 7.31% |
| 25 | Oleg Cassini | US | Brand | 42.29 | 7.21% |
| 26 | LC Waikiki | TR | Retailer | 42.10 | 6.71% |
| 27 | Twist | TR | Brand | 41.98 | 6.11% |
| 28 | Setrms | TR | Brand & retailer | 40.36 | 7.41% |
| 29 | By Saygı | TR | Brand & retailer | 40.35 | 5.01% |
| 30 | Sagaza | TR | Brand & retailer | 37.81 | 6.61% |
| 31 | Iconas Line | TR | Brand | 37.31 | 6.11% |
| 32 | Machka | TR | Brand | 35.10 | 7.21% |
| 33 | Kayra | TR | Brand | 34.92 | 7.41% |
| 34 | Berneva | TR | Brand | 28.26 | 5.01% |
| 35 | DCEY | TR | Retailer | 20.92 | 5.31% |
4. Two Shelves: Origin-Constrained and Origin-Neutral Questions
One-sentence takeaway: On the 20 origin-constrained questions domestic entities take 100.00% of qualifying mentions and foreign entities 0.00%, and on the 20 origin-neutral questions they take 83.21% and 16.79% — with origin on file for all 2,485 and all 1,769 qualifying mentions respectively.
The 40 questions divide into 20 that ask for Turkish brands and 20 that do not. Each set has its own leaderboard, its own denominator and its own scale. Neither is the market on its own, so the two figures are always quoted together, each with the number of questions behind it.
4.1 Origin share on both cuts
| Origin-constrained | Origin-neutral | All questions (blended — not an origin cut) | |
|---|---|---|---|
| Questions | 20 | 20 | 40 |
| Responses | 499 | 500 | 999 |
| Qualified entities in the share (domestic / foreign / unknown) | 36 (36 / 0 / 0) | 30 (25 / 5 / 0) | 35 (32 / 3 / 0) |
| Qualified entities left out (marketplaces and publishers), with mentions | Trendyol (176), Modanisa (93), Hepsiburada (85) | Trendyol (248), Modanisa (95), Hepsiburada (93), Akakçe (25) | Trendyol (424), Modanisa (188), Hepsiburada (178) |
| Domestic share of qualifying mentions | 100.00% | 83.21% | 94.68% |
| Foreign share of qualifying mentions | 0.00% | 16.79% | 5.32% |
| Qualifying mentions with origin on file | 2,485 of 2,485 | 1,769 of 1,769 | 4,133 of 4,133 |
| Sanity check — domestic share, all tracked entities | 99.64% | 83.28% | 92.59% |
| Sanity check — foreign share, all tracked entities | 0.36% | 16.72% | 7.41% |
| All-entity mentions with origin on file | 3,842 of 3,973 | 2,906 of 2,983 | 6,748 of 6,956 |
Origin-constrained (20 questions, 499 responses): domestic entities take 100.00% of qualifying mentions and foreign entities 0.00%, with origin on file for 2,485 of 2,485 qualifying mentions. All 36 qualified entities in the share are domestic.
Origin-neutral (20 questions, 500 responses): domestic entities take 83.21% and foreign entities 16.79%, with origin on file for 1,769 of 1,769. Of the 30 qualified entities in the share, 25 are domestic and 5 foreign.
The assistants read “yerli” as a constraint and answer it with domestic names only. Without it, foreign entities enter the shelf — Mango, H&M, Zara, Oleg Cassini and Pronovias — but domestic entities still take most of the qualifying mentions, because the four most-named entities on the neutral shelf that are not marketplaces — Carmen, Abiyefon, ADL and The ZKS — are domestic.
The sanity check points the same way. Across every tracked entity, qualified or not, domestic entities take 99.64% and foreign entities 0.36% of mentions on the constrained questions (origin on file for 3,842 of 3,973), and 83.28% and 16.72% on the neutral questions (2,906 of 2,983). On the constrained questions the long tail carries a sliver of foreign names that the shelf does not; on the neutral questions the two bases are close.
Left out of every share, and still ranked. On the constrained cut, three qualifying marketplaces are left out of the share: Trendyol (176 mentions), Modanisa (93) and Hepsiburada (85). On the neutral cut, four entities are: the same three marketplaces, Trendyol (248), Modanisa (95) and Hepsiburada (93), and the publisher Akakçe (25). Across all tracked entities, 14 marketplaces and publishers with 416 mentions are left out on the constrained cut, and 17 with 559 on the neutral cut. Every qualifying entity left out is domestic, so including them would raise the domestic share for a reason that has nothing to do with brands. They still rank in every leaderboard.
4.2 The origin-constrained shelf
39 entities qualify on the 20 origin-constrained questions (499 responses). Thirty-six enter the origin share, all domestic, and the three marketplaces are left out of it.
Carmen leads at 93.20, named in 46.89% of these responses (234 of 499), the highest mention rate on the cut. It scores below 100 because the cut’s highest MRR is not Carmen’s but Raisa Vanessa’s, 0.6311. The ZKS is second at 75.96 on 31.06% (155 of 499). Trendyol is third at 74.85: named more often than The ZKS, in 35.27% (176 of 499), but later, with an MRR of 0.3366 against 0.4449. AlfaBeta (72.93, 34.27%) and ADL (70.34, 30.06%) complete the top five.
This shelf brings smaller domestic names forward on position. Raisa Vanessa ranks 17th on 5.61% (28 of 499), named by 2 of 5 models, with the cut’s highest MRR. Eyyo Atelier ranks 15th on 6.81% (34 of 499) with an MRR of 0.3102, named by all five models, and Eser Giyim 20th on 7.01% (35 of 499). Kenzel and Cengiz Aktürk are named equally often, in 16.63% (83 of 499) each, and rank 13th and 14th. Near the foot of the board, Berneva, Machka, Iconas Line and Exquise each rest on one or two models (Chapter 11). Ranks built mainly on position, or on a single model, should be read cautiously.
No foreign entity qualifies on this cut. Foreign entities are named on the constrained questions — they take 0.36% of all-entity mentions with origin on file — but none often enough to reach the 5% threshold.
4.3 The origin-neutral shelf
| # | Entity | Origin | Type | Score | Mention rate | Mentions / responses | Breadth |
|---|---|---|---|---|---|---|---|
| 1 | Trendyol | TR | Marketplace | 100.00 | 49.60% | 248 / 500 | 5 of 5 |
| 2 | Carmen | TR | Brand & retailer | 71.91 | 30.00% | 150 / 500 | 5 of 5 |
| 3 | The ZKS | TR | Brand | 65.53 | 20.60% | 103 / 500 | 5 of 5 |
| 4 | Beymen | TR | Retailer | 63.32 | 13.60% | 68 / 500 | 5 of 5 |
| 5 | Abiyefon | TR | Brand & retailer | 63.25 | 22.00% | 110 / 500 | 5 of 5 |
| 6 | ADL | TR | Brand | 60.67 | 22.00% | 110 / 500 | 5 of 5 |
| 7 | Koton | TR | Brand | 59.03 | 19.60% | 98 / 500 | 5 of 5 |
| 8 | Modanisa | TR | Marketplace | 56.37 | 19.00% | 95 / 500 | 5 of 5 |
| 9 | Tarık Ediz | TR | Brand & retailer | 55.07 | 12.40% | 62 / 500 | 5 of 5 |
| 10 | Mango | ES | Brand | 54.26 | 13.00% | 65 / 500 | 5 of 5 |
| 11 | Hepsiburada | TR | Marketplace | 53.72 | 18.60% | 93 / 500 | 4 of 5 |
| 12 | Vakko | TR | Retailer | 53.18 | 14.00% | 70 / 500 | 5 of 5 |
| 13 | H&M | SE | Brand | 53.10 | 8.80% | 44 / 500 | 5 of 5 |
| 14 | Zara | ES | Brand | 51.31 | 16.40% | 82 / 500 | 5 of 5 |
| 15 | Sefamerve | TR | Retailer | 51.06 | 13.40% | 67 / 500 | 5 of 5 |
| 16 | AlfaBeta | TR | Brand & retailer | 47.63 | 14.20% | 71 / 500 | 5 of 5 |
| 17 | GIZIA | TR | Brand | 47.31 | 8.60% | 43 / 500 | 5 of 5 |
| 18 | Roman | TR | Brand | 44.61 | 9.40% | 47 / 500 | 5 of 5 |
| 19 | Oleg Cassini | US | Brand | 44.17 | 13.60% | 68 / 500 | 3 of 5 |
| 20 | Pronovias | ES | Brand & retailer | 43.38 | 7.60% | 38 / 500 | 5 of 5 |
| 21 | Lafaba | TR | Brand | 43.09 | 7.80% | 39 / 500 | 5 of 5 |
| 22 | EKOL | TR | Brand & retailer | 41.97 | 5.00% | 25 / 500 | 4 of 5 |
| 23 | Setre | TR | Brand & retailer | 41.57 | 9.40% | 47 / 500 | 4 of 5 |
| 24 | Kenzel | TR | Brand | 41.00 | 7.40% | 37 / 500 | 5 of 5 |
| 25 | Twist | TR | Brand | 39.89 | 7.80% | 39 / 500 | 5 of 5 |
| 26 | Nocturne | TR | Brand | 39.54 | 9.80% | 49 / 500 | 4 of 5 |
| 27 | LC Waikiki | TR | Retailer | 39.14 | 9.00% | 45 / 500 | 4 of 5 |
| 28 | İpekyol | TR | Brand | 38.18 | 8.80% | 44 / 500 | 3 of 5 |
| 29 | Cengiz Aktürk | TR | Brand & retailer | 37.77 | 5.60% | 28 / 500 | 4 of 5 |
| 30 | Rengin | TR | Brand | 33.69 | 5.40% | 27 / 500 | 4 of 5 |
| 31 | Machka | TR | Brand | 31.20 | 6.80% | 34 / 500 | 3 of 5 |
| 32 | Akakçe | TR | Publisher | 29.13 | 5.00% | 25 / 500 | 3 of 5 |
| 33 | Network | TR | Brand | 27.68 | 5.00% | 25 / 500 | 3 of 5 |
| 34 | DCEY | TR | Retailer | 21.18 | 6.80% | 34 / 500 | 2 of 5 |
Trendyol leads this shelf at 100.00, named in 49.60% of these responses (248 of 500) with an MRR of 0.5760. Carmen (71.91, 30.00%) and The ZKS (65.53, 20.60%) follow. Beymen is fourth at 63.32 on 13.60% (68 of 500), with an MRR of 0.4989; Abiyefon and ADL, fifth and sixth, are each named in 22.00% (110 of 500) and rank below it.
Five foreign entities qualify. Mango is 10th at 54.26 on 13.00% (65 of 500). H&M, Swedish, is 13th at 53.10 on 8.80% (44 of 500), with an MRR of 0.3861. Zara is 14th at 51.31, named more often than either, in 16.40% (82 of 500), but later (MRR 0.2195). Oleg Cassini is 19th at 44.17 on 13.60% (68 of 500), named by 3 of 5 models, and Pronovias, Spanish, 20th at 43.38 on 7.60% (38 of 500). Oleg Cassini’s visibility leans on one model: on the general cut, 57 of its 72 mentions (79.17%) come from Gemini 3.6 Flash, which joined the panel this edition (Chapter 5).
Modanisa (8th, 19.00%) and Hepsiburada (11th, 18.60%) are the other marketplaces on this shelf. Akakçe, a publisher, qualifies here too (32nd); like the marketplaces, it ranks and is left out of the share.
4.4 What this chapter does not compare
No origin share here is compared with July’s, because July assigned origin by a different rule (section 2.7) and published its origin-neutral split on its own basis. The constrained figure is never read on its own: 100.00% domestic and 0.00% foreign on 20 questions that ask for Turkish brands describes the question as much as the market, and it stands beside 83.21% domestic and 16.79% foreign on the 20 that do not. The all-questions column in the table above blends the two scopes and is not an origin measurement. And the question-type cuts in Chapter 6 each blend constrained and neutral questions, so their counts of domestic qualifiers are not origin measurements either.
5. Differences Between Models
One-sentence takeaway: Carmen is in every model’s top three, but the models differ sharply in length — Gemini 3.6 Flash names 12.38 entities per answer and Claude Haiku 4.5 4.44 — and 11 qualifying ranks draw at least 60% of their general-cut mentions from a single model.
5.1 Per-model behavior
| Model | Responses | Web-grounded | Citations | Distinct entities | Entities per answer | Most-named entity | Its rate | In July's panel |
|---|---|---|---|---|---|---|---|---|
| Claude Haiku 4.5 | 200 | 100.00% | 1,000 | 148 | 4.44 | The ZKS (TR) | 27.00% | Yes |
| Gemini 3.6 Flash | 200 | 20.00% | 388 | 170 | 12.38 | Carmen (TR) | 47.50% | No |
| ChatGPT 5.6 Luna Pro | 199 | 97.99% | 937 | 159 | 4.91 | Trendyol (TR) | 35.68% | No |
| Perplexity Sonar | 200 | 100.00% | 5,930 | 160 | 6.54 | Carmen (TR) | 37.50% | Yes |
| Grok 4.3 | 200 | 100.00% | 1,051 | 171 | 11.40 | Trendyol (TR) | 87.50% | Yes |
The leaders are shared; their order is not. Trendyol is the most-named entity for Grok 4.3 (87.50% of its 200 answers) and ChatGPT 5.6 Luna Pro (35.68% of its 199). Carmen is the most-named for Gemini 3.6 Flash (47.50%) and Perplexity Sonar (37.50%), and The ZKS for Claude Haiku 4.5 (27.00%). Carmen is in the top three of all five models.
Answer length is the largest difference. Gemini 3.6 Flash names 12.38 entities per answer and Grok 4.3 11.40, and they reach the widest repertoires, 170 and 171 distinct entities. Perplexity Sonar names 6.54 per answer (160 distinct). ChatGPT 5.6 Luna Pro (4.91, 159 distinct) and Claude Haiku 4.5 (4.44, 148 distinct) give short, selective answers. The two models that joined this edition sit at opposite ends: Gemini 3.6 Flash writes the longest lists on the panel, ChatGPT 5.6 Luna Pro some of the shortest. For an entity at the edge of visibility, being named at all depends heavily on which assistant answers.
Entities named per answer, by model
5.2 Each model’s most-named entities
| # | Claude Haiku 4.5 | Gemini 3.6 Flash | ChatGPT 5.6 Luna Pro | Perplexity Sonar | Grok 4.3 |
|---|---|---|---|---|---|
| 1 | The ZKS 27.00% | Carmen 47.50% | Trendyol 35.68% | Carmen 37.50% | Trendyol 87.50% |
| 2 | Carmen 19.00% | İpekyol 42.50% | The ZKS 31.16% | Trendyol 35.50% | Carmen 59.50% |
| 3 | Abiyefon 14.50% | Trendyol 40.00% | Carmen 28.64% | ADL 28.50% | Hepsiburada 55.50% |
| 4 | Trendyol 13.50% | AlfaBeta 38.00% | Berneva 25.13% | Koton 27.50% | The ZKS 52.50% |
| 5 | ADL 11.50% | Tarık Ediz 36.50% | ADL 15.58% | Abiyefon 23.00% | Sefamerve 51.00% |
| 6 | AlfaBeta 11.50% | Modanisa 34.00% | Tuba Ergin 12.56% | AlfaBeta 20.50% | Abiyefon 42.00% |
| 7 | Hepsiburada 9.00% | ADL 33.00% | Modanisa 12.06% | LC Waikiki 19.50% | ADL 41.50% |
| 8 | Kenzel 8.50% | Nocturne 32.50% | Coupe Official 10.55% | Sefamerve 19.50% | AlfaBeta 40.50% |
| 9 | Koton 7.50% | Machka 31.00% | Koton 10.55% | The ZKS 16.00% | Modanisa 38.50% |
| 10 | GIZIA 7.50% | Setrms 30.00% | AlfaBeta 10.55% | Hepsiburada 15.00% | Koton 33.50% |
The models agree on a core: Trendyol, Carmen, ADL and AlfaBeta are in every model’s top ten. They differ beneath it. Grok 4.3 has three marketplaces in its top ten — Trendyol first (87.50%), Hepsiburada third (55.50%) and Modanisa ninth (38.50%) — and the retailer Sefamerve fifth (51.00%). Gemini 3.6 Flash has İpekyol second (42.50%), Tarık Ediz fifth (36.50%), and Nocturne, Machka and Setrms eighth to tenth (32.50%, 31.00% and 30.00%); none of the five is in any other model’s top ten, and The ZKS is not in Gemini 3.6 Flash’s. ChatGPT 5.6 Luna Pro has Berneva fourth (25.13%), Tuba Ergin sixth (12.56%) and Coupe Official eighth (10.55%), none of them in another model’s top ten. Claude Haiku 4.5 is the most compact: its eighth entity, Kenzel, is named in 8.50% of its answers and its tenth, GIZIA, in 7.50%. Perplexity Sonar is the only model with LC Waikiki in its top ten (seventh, 19.50%).
5.3 One model’s habit
Eleven qualifying entities on the general cut draw at least three in five of their mentions from a single model:
| Entity | General rank | Mentions, all models | From one model | That model's share | That model's rate | Models naming |
|---|---|---|---|---|---|---|
| Berneva | 37 | 50 | ChatGPT 5.6 Luna Pro (50) | 100.00% | 25.13% of 199 | 1 |
| DCEY | 38 | 53 | Gemini 3.6 Flash (50) | 94.34% | 25.00% of 200 | 2 |
| Sagaza | 33 | 66 | Gemini 3.6 Flash (59) | 89.39% | 29.50% of 200 | 2 |
| Iconas Line | 34 | 61 | Grok 4.3 (54) | 88.52% | 27.00% of 200 | 4 |
| Machka | 35 | 72 | Gemini 3.6 Flash (62) | 86.11% | 31.00% of 200 | 3 |
| Setrms | 31 | 74 | Gemini 3.6 Flash (60) | 81.08% | 30.00% of 200 | 4 |
| Kayra | 36 | 74 | Gemini 3.6 Flash (59) | 79.73% | 29.50% of 200 | 3 |
| Oleg Cassini | 28 | 72 | Gemini 3.6 Flash (57) | 79.17% | 28.50% of 200 | 3 |
| İpekyol | 20 | 109 | Gemini 3.6 Flash (85) | 77.98% | 42.50% of 200 | 4 |
| Setre | 19 | 96 | Grok 4.3 (60) | 62.50% | 30.00% of 200 | 5 |
| Hepsiburada | 13 | 178 | Grok 4.3 (111) | 62.36% | 55.50% of 200 | 4 |
Berneva is ChatGPT 5.6 Luna Pro’s
Berneva qualifies on the general cut — 37th, on 5.01% of responses (50 of 999) — on mentions from one model only. All 50 come from ChatGPT 5.6 Luna Pro, which names it in 25.13% of its 199 answers and ranks it fourth; no other model names it at all. ChatGPT 5.6 Luna Pro joined the panel this edition. Berneva’s rank describes one assistant’s answers, not a consensus across the panel.
DCEY, Sagaza, Machka, Setrms, Kayra, Oleg Cassini and İpekyol lean on Gemini 3.6 Flash
Gemini 3.6 Flash supplies 50 of DCEY’s 53 general-cut mentions (94.34%), 59 of Sagaza’s 66 (89.39%), 62 of Machka’s 72 (86.11%), 60 of Setrms’s 74 (81.08%), 59 of Kayra’s 74 (79.73%), 57 of Oleg Cassini’s 72 (79.17%) and 85 of İpekyol’s 109 (77.98%). Gemini 3.6 Flash also joined this edition, and it writes the longest lists on the panel, 12.38 entities per answer; a long answer gives more names a chance to be counted at all. All of them sit in the lower half of the general board, from İpekyol (20th) to DCEY (38th). Oleg Cassini is one of the three foreign qualifiers on the general cut and one of the five on the origin-neutral cut (Chapter 4), so part of the foreign presence on the shelf is this one model’s.
Iconas Line, Setre and Hepsiburada lean on Grok 4.3
Grok 4.3 supplies 54 of Iconas Line’s 61 general-cut mentions (88.52%), 60 of Setre’s 96 (62.50%) and 111 of Hepsiburada’s 178 (62.36%), naming them in 27.00%, 30.00% and 55.50% of its 200 answers. Grok 4.3 answered in both editions. Hepsiburada, 13th, is a marketplace: its rank describes one assistant’s habit of pointing to where to buy as much as a consensus across the panel.
5.4 What only one model names
Every model names entities no other model names at all: 57 for Gemini 3.6 Flash, 42 for Grok 4.3, 33 for ChatGPT 5.6 Luna Pro, 25 for Claude Haiku 4.5 and 24 for Perplexity Sonar.
| Model | Entities named by this model alone | Examples |
|---|---|---|
| ChatGPT 5.6 Luna Pro | 33 | Berneva, Coupe Official, So Lady, Farika, Miss Diamond, Pionto |
| Claude Haiku 4.5 | 25 | Didem Soydan, GILLA'S QUEEN, Desigual, Just Cavalli, Karl Lagerfeld, United Colors of Benetton |
| Gemini 3.6 Flash | 57 | Tuvanam, Ju Istanbul, Touche Privé, Exquise, Tuğba, Gigii's |
| Grok 4.3 | 42 | Mayabella, Beyza Giyim, Beyza Online, Dressland, Davetten Davete, Milagron |
| Perplexity Sonar | 24 | Ceylan Otantik, Civil, Laranor, Jeyan Gedik, İmajButik, HESNA |
6. Question-Type Ownership
One-sentence takeaway: Carmen leads the discovery questions and Trendyol leads the attribute and use-case questions; discovery qualifies 53 entities against 34 for attribute and 32 for use-case questions, and the specialists below the leaders change with the type.
| Question type | Questions | Responses | Distinct entities | Qualified | Domestic qualifiers (blended) | Average score | Leader | Highest mention rate | Highest MRR |
|---|---|---|---|---|---|---|---|---|---|
| Discovery | 14 | 350 | 291 | 53 | 48 | 43.80 | Carmen (88.84) | 48.29% (Carmen) | 0.6988 (Raisa Vanessa) |
| Attribute | 13 | 325 | 220 | 34 | 33 | 48.40 | Trendyol (90.85) | 43.38% (Trendyol) | 0.6133 (Faik Sönmez) |
| Use case | 13 | 324 | 226 | 32 | 31 | 49.13 | Trendyol (100.00) | 38.58% (Trendyol) | 0.5269 (Trendyol) |
| # | Entity | Score | Rate | MRR |
|---|---|---|---|---|
| 1 | Carmen | 88.84 | 48.29% | 0.4388 |
| 2 | Trendyol | 87.76 | 45.14% | 0.4819 |
| 3 | ADL | 79.67 | 43.43% | 0.3307 |
| 4 | Tarık Ediz | 70.17 | 30.29% | 0.3948 |
| 5 | AlfaBeta | 69.43 | 37.14% | 0.2287 |
| # | Entity | Score | Rate | MRR |
|---|---|---|---|---|
| 1 | Trendyol | 90.85 | 43.38% | 0.4262 |
| 2 | Carmen | 85.94 | 39.69% | 0.4041 |
| 3 | The ZKS | 84.65 | 30.77% | 0.5669 |
| 4 | Abiyefon | 77.80 | 30.77% | 0.4268 |
| 5 | Faik Sönmez | 61.38 | 6.15% | 0.6133 |
| # | Entity | Score | Rate | MRR |
|---|---|---|---|---|
| 1 | Trendyol | 100.00 | 38.58% | 0.5269 |
| 2 | Carmen | 85.52 | 26.54% | 0.5192 |
| 3 | Koton | 77.85 | 22.53% | 0.4667 |
| 4 | The ZKS | 77.53 | 22.84% | 0.4546 |
| 5 | AlfaBeta | 67.43 | 14.20% | 0.4544 |
Discovery — “which brand?” or “where to buy?” — spreads visibility widest: 291 distinct entities, 53 qualifiers and the lowest average score of the three, 43.80. Carmen leads at 88.84 and is named most, in 48.29% of these responses (169 of 350). Trendyol is second at 87.76 on 45.14% (158 of 350), and ADL third at 79.67 on 43.43% (152 of 350). The cut’s highest MRR is Raisa Vanessa’s, 0.6988, on mentions from two models (27 mentions). Berneva, Tuvanam, Iconas Line, Ju Istanbul and Touche Privé each qualify on this cut on mentions from a single model.
Attribute — a specific property — draws 220 distinct entities and 34 qualifiers, with an average score of 48.40. Trendyol leads at 90.85 and is named most, in 43.38% of these responses (141 of 325); Carmen is second at 85.94 on 39.69%. The ZKS and Abiyefon are named equally often, in 30.77% (100 of 325) each, and The ZKS ranks third on an MRR of 0.5669 against Abiyefon’s 0.4268. Faik Sönmez is fifth at 61.38 on a 6.15% mention rate (20 of 325), carried by the cut’s highest MRR, 0.6133, and named by all five models. Oleg Cassini, Kayra, Deniz Butik, Setrms and Alvina each qualify here on two models.
Use case — an occasion or need — draws 226 distinct entities and 32 qualifiers, with an average score of 49.13. Trendyol leads at 100.00, named in 38.58% of these responses (125 of 324) and holding the cut’s highest MRR, 0.5269; Carmen is second at 85.52 on 26.54%. Koton is third at 77.85 on 22.53% (73 of 324), just ahead of The ZKS (77.53), which is named slightly more often, in 22.84%, but later: MRR 0.4546 against Koton’s 0.4667. Nocturne, Sagaza, Tarık Ediz and Iconas Line each qualify here on two models.
The three types rest on 350, 325 and 324 responses. The differences between them are not origin findings: discovery splits 7 origin-neutral and 7 origin-constrained questions, attribute 7 and 6, and use case 6 and 7, and Chapter 4 is where origin is measured.
7. Attribute Axes
One-sentence takeaway: No attribute axis can be published this edition: no entity has a product category or audience on file, and no subcategory was declared before collection, which this series requires before publishing one as a cut.
An attribute chapter ranks entities by something they are — the product categories they make, or the audience they serve — rather than by the questions asked. This edition’s data carries no such attribute: none of the 393 tracked entities has a product category on file, and no audience attribute is recorded. A product-shelf chapter cannot be written from this export, and this report does not substitute an estimate for one.
No question in this study records a subcategory either, and under this series’ rules a subcategory is published as a cut only when the study declares it before collection. This study did not. July’s questions carried no topical subcategories, and July read the three question types instead; that reading continues, with full leaderboards, in Chapter 6.
8. What Changed Since July
One-sentence takeaway: On the same 40 questions, against July’s answers counted the way this edition counts, 74 of the 198 entities named in both editions rose and 111 fell — The ZKS gained 19.01 points and Iconas Line lost 8.54 — with two of the five models replaced between editions, which no figure here corrects for.
8.1 How this comparison is built
This chapter compares raw mention rate on the general cut; the general board’s rank and score change since July are shown in Chapter 3. All 40 questions were asked in both editions, so the change on the shared questions is the plain difference between the two levels: July’s 997 responses against October’s 999.
July’s levels are recounts. Every July figure in this chapter is July’s answers, counted the way this edition counts — re-read with this edition’s brand registry so both sides count a brand the same way. They are not the figures July’s page printed. July’s page counted names under a different rule (section 2.6), and the registry has changed since July, so any entity’s July level here can differ from July’s page. No figure from July’s page is quoted in this chapter.
The panel is the confound, and it is declared, not corrected. Claude Haiku 4.5, Perplexity Sonar and Grok 4.3 answered in both editions. Gemini 2.5 Flash-Lite and GPT-4o-mini answered in July only; Gemini 3.6 Flash and ChatGPT 5.6 Luna Pro answered in October only. Rebasing adjusts for the question set, which did not change; nothing adjusts for the panel. Every number below carries that confound: the direction of a large move is better supported than its exact size.
Origin in this chapter is this edition’s coding, applied to both editions. July’s own origin rule is not re-read, so the domestic and foreign counts below are one rule applied twice, not an origin share compared across the rule change.
| Summary | |
|---|---|
| Questions | 40 shared; 0 added; 0 retired |
| Responses | July 997 (recounted); October 999 |
| Models in both editions | Claude Haiku 4.5, Perplexity Sonar, Grok 4.3 |
| Joined / left | Gemini 3.6 Flash, ChatGPT 5.6 Luna Pro / Gemini 2.5 Flash-Lite, GPT-4o-mini |
| Entities named in both editions | 198: rose 74, fell 111, unchanged 13 |
| — domestic: held / rose | 157 / 67 |
| — foreign: held / rose | 33 / 6 |
| — no origin on file: held | 8 |
| Named only in October / only in July | 195 / 96 |
| Largest gain | The ZKS (+19.01 points) |
| Largest fall | Iconas Line (-8.54 points) |
Of the 198 entities named in both editions, 74 rose, 111 fell and 13 were unchanged. Domestic entities account for 157 of the 198, of which 67 rose; foreign entities for 33, of which 6 rose; 8 have no origin on file and are neither. A further 195 entities were named only in October’s answers and 96 only in July’s. Those entities are not in the tables below, which rank only entities named in both editions.
8.2 The movers
| Entity | Origin | July (recounted) | October | Δ points |
|---|---|---|---|---|
| The ZKS | TR | 6.82% (68 / 997) | 25.83% (258 / 999) | +19.01 |
| Nocturne ⚠ | TR | 0.20% (2 / 997) | 10.91% (109 / 999) | +10.71 |
| İpekyol | TR | 1.10% (11 / 997) | 10.91% (109 / 999) | +9.81 |
| Vakko | TR | 1.40% (14 / 997) | 10.71% (107 / 999) | +9.31 |
| Tarık Ediz | TR | 7.22% (72 / 997) | 15.62% (156 / 999) | +8.39 |
| Sefamerve | TR | 12.94% (129 / 997) | 21.32% (213 / 999) | +8.38 |
| Kayra ⚠ | TR | 0.30% (3 / 997) | 7.41% (74 / 999) | +7.11 |
| Abiyefon | TR | 16.25% (162 / 997) | 23.32% (233 / 999) | +7.07 |
| Armine ⚠ | TR | 0.50% (5 / 997) | 7.31% (73 / 999) | +6.81 |
| Modanisa | TR | 12.14% (121 / 997) | 18.82% (188 / 999) | +6.68 |
| Entity | Origin | July (recounted) | October | Δ points |
|---|---|---|---|---|
| Iconas Line | TR | 14.64% (146 / 997) | 6.11% (61 / 999) | -8.54 |
| ModaSelvim | TR | 11.63% (116 / 997) | 6.31% (63 / 999) | -5.33 |
| Pronovias | ES | 9.03% (90 / 997) | 3.90% (39 / 999) | -5.12 |
| VznModa | TR | 6.62% (66 / 997) | 1.60% (16 / 999) | -5.02 |
| Hülya Keser | TR | 7.52% (75 / 997) | 2.50% (25 / 999) | -5.02 |
| Lacrima | TR | 8.02% (80 / 997) | 3.20% (32 / 999) | -4.82 |
| EKOL | TR | 10.13% (101 / 997) | 6.51% (65 / 999) | -3.62 |
| Yeşim Gelinlik ⚠ | TR | 4.21% (42 / 997) | 0.90% (9 / 999) | -3.31 |
| Boyner | TR | 5.62% (56 / 997) | 2.30% (23 / 999) | -3.31 |
| Modazehrada ⚠ | TR | 3.81% (38 / 997) | 0.70% (7 / 999) | -3.11 |
Gains. The ZKS moved from 6.82% of July’s recounted answers (68 of 997) to 25.83% (258 of 999), +19.01 points, and is third on the general board this edition. İpekyol moved from 1.10% (11 of 997) to 10.91% (109 of 999), +9.81, and Vakko from 1.40% (14 of 997) to 10.71% (107 of 999), +9.31. Tarık Ediz (+8.39, from 7.22% to 15.62%), Sefamerve (+8.38, from 12.94% to 21.32%), Abiyefon (+7.07, from 16.25% to 23.32%) and the marketplace Modanisa (+6.68, from 12.14% to 18.82%) rose too. Three of the ten largest gains rest on too few July mentions to read as more than direction: Nocturne ⚠ (+10.71, from 2 of 997 to 109 of 999), Kayra ⚠ (+7.11, from 3 to 74) and Armine ⚠ (+6.81, from 5 to 73). All ten gainers are domestic.
Falls. Iconas Line fell 8.54 points, from 14.64% (146 of 997) to 6.11% (61 of 999), and still qualifies, 34th on the general board. ModaSelvim (-5.33, from 11.63% to 6.31%) and EKOL (-3.62, from 10.13% to 6.51%) also fell and still qualify. Pronovias, Spanish, fell 5.12 points, from 9.03% to 3.90%, and is below the general-cut threshold this edition; it does qualify on the origin-neutral cut, 20th (Chapter 4). VznModa (-5.02), Hülya Keser (-5.02), Lacrima (-4.82) and the retailer Boyner (-3.31, from 5.62% to 2.30%) also sit below the threshold this edition. Yeşim Gelinlik ⚠ (-3.31, to 9 of 999) and Modazehrada ⚠ (-3.11, to 7 of 999) rest on too few mentions this edition to read as more than direction.
8.3 Where the gains sit
This report publishes no shared-model comparison and carries no per-model July figures, so it cannot say how much of any gain the panel change accounts for. What it can show is where each mover is named now.
The ZKS’s current visibility does not sit only in the models that joined. It is Claude Haiku 4.5’s most-named entity (27.00%), fourth for Grok 4.3 (52.50%) and ninth for Perplexity Sonar (16.00%) — all three answered in both editions — as well as second for ChatGPT 5.6 Luna Pro (31.16%), which joined. It is not in Gemini 3.6 Flash’s top ten. Abiyefon and Sefamerve have similar profiles: Abiyefon is third for Claude Haiku 4.5 (14.50%), fifth for Perplexity Sonar (23.00%) and sixth for Grok 4.3 (42.00%), and Sefamerve fifth for Grok 4.3 (51.00%) and eighth for Perplexity Sonar (19.50%); neither is in the top ten of a joining model. Whether any of them rose within the carried models cannot be read from this report.
Nocturne ⚠, İpekyol, Kayra ⚠ and Tarık Ediz lean on Gemini 3.6 Flash. Nocturne is eighth in Gemini 3.6 Flash’s answers (32.50%) and in no other model’s top ten. Gemini 3.6 Flash supplies 85 of İpekyol’s 109 general-cut mentions and 59 of Kayra’s 74 (Chapter 5). Tarık Ediz is fifth for Gemini 3.6 Flash (36.50%) and in no other model’s top ten. Gemini 3.6 Flash joined the panel this edition, so for these entities the panel change cannot be ruled out as a large part of the gain.
Modanisa is spread across both. It is in the top ten of one carried model, Grok 4.3 (38.50%), and of both joining models, Gemini 3.6 Flash (34.00%) and ChatGPT 5.6 Luna Pro (12.06%). Vakko and Armine ⚠ are in no model’s top ten, so this report cannot say where their gains sit.
Iconas Line’s remaining mentions are mostly Grok 4.3’s — 54 of its 61 (88.52%) — and Grok 4.3 answered in both editions. This report cannot say whether the fall happened within Grok 4.3 or reflects the models that left the panel.
9. The Discovery Ecosystem: Where AI Learns About Brands
One-sentence takeaway: The most widely cited of 213 domains is trendyol.com, in 84 of 200 question-model cells (42.00%), followed by carmen.com.tr and thezks.com; one model’s citations are capped and another’s grounding is partial, so the map is uneven by model.
When an assistant answers with web search, it draws on whatever is available to it at that moment. The cited addresses were reduced to domains and counted by question-model cell — one cell is one question put to one model, holding that question’s repeats — never by response. Herm.io’s own domain was removed before any of these counts. Three units appear in this chapter, and they are not interchangeable:
- Citations recorded — 9,306. The citation entries the models returned.
- Link rows — 2,907. Attribution rows, one per model, entity, domain and attribution rule; one mention can be attributed to several domains.
- Mentions attributed — 7,738. Those rows’ mention counts, summed. Because one mention can sit on several rows, this is not a count of distinct mentions and does not compare with any mention rate’s numerator.
Across all answers, 213 domains were cited, 70 of them in a single cell only.
9.1 The most-cited domains
| # | Domain | Cells citing | Share of cells | Mentions attributed | Owner |
|---|---|---|---|---|---|
| 1 | trendyol.com | 84 | 42.00% | 888 | Trendyol |
| 2 | carmen.com.tr | 69 | 34.50% | 432 | Carmen |
| 3 | thezks.com | 59 | 29.50% | 537 | The ZKS |
| 4 | abiyefon.com | 51 | 25.50% | 218 | Abiyefon |
| 5 | alfa-beta.com.tr | 43 | 21.50% | 218 | AlfaBeta |
| 6 | adl.com.tr | 41 | 20.50% | 239 | ADL |
| 7 | sefamerve.com | 30 | 15.00% | 178 | Sefamerve |
| 8 | modanisa.com | 28 | 14.00% | 163 | Modanisa |
| 9 | hepsiburada.com | 26 | 13.00% | 126 | Hepsiburada |
| 10 | koton.com | 25 | 12.50% | 236 | Koton |
| 11 | kenzel.com.tr | 22 | 11.00% | 80 | Kenzel |
| 12 | gizia.com | 19 | 9.50% | 122 | GIZIA |
| 13 | cengizakturk.com | 19 | 9.50% | 63 | Cengiz Aktürk |
| 14 | roman.com.tr | 18 | 9.00% | 71 | Roman |
| 15 | lcw.com | 17 | 8.50% | 96 | LC Waikiki |
trendyol.com is cited in 84 of the 200 cells (42.00%), followed by carmen.com.tr (69 cells, 34.50%), thezks.com (59, 29.50%), abiyefon.com (51, 25.50%), alfa-beta.com.tr (43, 21.50%) and adl.com.tr (41, 20.50%). These six are the domains of the general board’s top six entities. Every domain in the table has an owner among the tracked entities: three are marketplaces — trendyol.com, modanisa.com and hepsiburada.com — and the rest are the own sites of brands and retailers. thezks.com carries more attributed mentions than carmen.com.tr, 537 against 432, from fewer cells — a reminder that attribution rows count every entity a cited page is credited with.
A cited address is a source, not a mention. carmen.com.tr counts here as a domain the models read; the same address written as a name inside an answer is not counted as Carmen (section 2.4).
9.2 Entities’ own sites
184 of the 213 domains are the own site of some tracked entity. Tracked is not ranked: marketplaces, publishers and entities that qualify in no cut have sites too, so this count does not reconcile with the leaderboard, and it should not be read as brands citing themselves. Rows on those domains account for 4,645 of the 7,738 attributed mentions — attribution-row units, not distinct mentions.
That figure is understated. Seven cited domains look like a tracked entity’s own site but are not attributed to it, because the registry holds a different website: aytimurtekstil.com (Aytimur Tekstil), erenabiye.com (Eren Abiye; registry erenabiye.com.tr), esergiyim.com.tr (Eser Giyim), matikabiyem.com (Matik Abiyem), nihan.com.tr (Nihan; registry nihangiyim.com), tubaergin.com.tr (Tuba Ergin; registry tubaergin.com) and vogue.com.tr (Vogue; registry vogue.com). They are therefore not counted among own sites.
In 61 rows, the rule that assigns a mention to the entity’s own domain and the rule that assigns it by position in the answer disagreed about which citation the mention belongs to. Those conflicts are surfaced in the data, not resolved.
9.3 Why the map is uneven
The models do not contribute equally. Claude Haiku 4.5 returned exactly 5 citations on every one of its 200 answers — 1,000 in all — so its domain counts stop at a ceiling and do not compare with an uncapped model’s. Perplexity Sonar alone returned 5,930 of the 9,306 citations. Gemini 3.6 Flash returned 388 and grounded 20.00% of its answers. A domain’s place in this table therefore reflects which models cite at length as much as what any one model reads.
This edition carries no source-type classification of domains and no per-response citation counts, so July’s source-type mix has no counterpart here, and July’s counts of responses citing a domain are a different unit from the cells counted above. Nothing in this chapter tests whether being cited causes visibility or merely accompanies it.
10. What the Patterns Suggest
One-sentence takeaway: Six neutral readings, not recommendations: a marketplace sets the top of the board, the shelf stays mostly domestic on both question scopes, frequency and position are separate routes onto the board, the two joining models carry several ranks near the threshold, answer format decides who is seen in the long tail, and the top entities’ own domains are the most cited.
This chapter describes patterns in the data. It is not advice, and visibility remains separate from quality.
A marketplace sets the top of the board. Trendyol leads the general cut (100.00), the origin-neutral cut (100.00) and the attribute (90.85) and use-case (100.00) questions; Carmen leads the origin-constrained cut (93.20) and the discovery questions (88.84). Because Trendyol holds every maximum on the general cut, every general-cut score — including those on the brand-only board — is read against a marketplace.
The shelf stays mostly domestic on both question scopes. On the 20 origin-constrained questions domestic entities take 100.00% of qualifying mentions and foreign entities 0.00%; on the 20 origin-neutral questions, 83.21% and 16.79% — with origin on file for every qualifying mention on both. The assistants read “yerli” as a constraint; without it, five foreign entities qualify, the highest-placed of them Mango in 10th.
Frequency and position are separate routes onto the board. ADL is named in 26.03% of all responses and ranks fourth; The ZKS, in 25.83%, ranks third because it is named earlier. Beymen ranks tenth on 7.61% with an MRR of 0.4565. On narrower cuts the effect is sharper: Raisa Vanessa ranks 17th on the origin-constrained questions on 5.61% with an MRR of 0.6311, and Faik Sönmez fifth on the attribute questions on 6.15% with an MRR of 0.6133.
The two joining models carry several ranks near the threshold. Gemini 3.6 Flash supplies most of the general-cut mentions of DCEY, Sagaza, Machka, Setrms, Kayra, Oleg Cassini and İpekyol; ChatGPT 5.6 Luna Pro supplies all of Berneva’s. All of them sit between 20th and 38th on the general board. A change of panel moved the edge of this board more visibly than its top.
Answer format decides who is seen in the long tail. Gemini 3.6 Flash names 12.38 entities per answer, reaches 170 distinct entities and names 57 that no other model does; Claude Haiku 4.5 names 4.44 per answer, 148 distinct and 25 alone. For an entity at the edge of the board, whether it is seen depends on the assistant’s format as much as on the entity.
The top entities’ own domains are the most cited. trendyol.com, carmen.com.tr, thezks.com, abiyefon.com, alfa-beta.com.tr and adl.com.tr are the six most widely cited domains, in 84, 69, 59, 51, 43 and 41 of 200 cells, and they belong to the general board’s top six entities. This is an observed association between being cited and being named, not evidence that one causes the other.
11. Limitations and Notes
One-sentence takeaway: This is one market on one collection day, measured on a panel, a counting rule and an origin rule that all changed since July, with scope unresolved for every entity; read every figure with the notes below.
- The panel changed. Gemini 2.5 Flash-Lite and GPT-4o-mini were replaced by Gemini 3.6 Flash and ChatGPT 5.6 Luna Pro. Chapter 8 declares this confound beside every figure; it does not correct for it, and some movers — Nocturne ⚠, İpekyol and Kayra ⚠ among them — are named now mainly by Gemini 3.6 Flash, which joined.
- The counting rule changed. July’s page counted a validated brand name wherever it appeared in the answer text; this edition counts a brand or company only where a second language model lists it as named in the answer, not where it appears only in a web address. July’s levels in this report are therefore recounts of July’s answers, not the figures July printed.
- The registry changed. Brands have been renamed and registrations reconciled since July. The recount of July’s answers uses this edition’s registry.
- The origin rule changed. July required a brand to be founded and headquartered in Turkey; this edition uses brand home. No origin share is compared across the change.
- The brand-only view changed. July’s removed marketplaces and retailers; this edition’s removes marketplaces only, so the two views are different lists.
- Origin classification involves judgment. Brand home is a researched decision, and brands run from one country under an owner in another are genuinely arguable cases. 40 entities have no origin on file; they appear in every table and in no share.
- Scope was not resolved. No entity has a product category on file, so the study could not separate core evening-wear entities from adjacent ones, and all 393 are eligible to rank. An entity whose main business lies outside evening wear can rank if the assistants name it often enough.
- Untyped and provisional entities. Two entities, Didem Soydan and Armut.com, are both untyped and provisional; neither qualifies in any cut. Armut.com is a likely duplicate of the marketplace Armut, which is tracked separately.
- Market-suffix names. Two names carry a market suffix whose parent brand is not tracked in this edition, so nothing could be folded into a parent; each keeps its own origin.
- Brands written as web addresses. A brand written in an answer only as its web address is not counted as that brand: “Carmen.com.tr”, in 11 responses, is not counted as Carmen (section 2.4).
- One model’s answers have unknown grounding. 160 of Gemini 3.6 Flash’s 200 answers, and 4 of ChatGPT 5.6 Luna Pro’s 199 analysed answers, returned no grounding metadata. They are not memory answers, and nothing in this report treats them as such.
- One model’s citations are capped. Claude Haiku 4.5 returned exactly 5 citations per answer, so its source counts are truncated (Chapter 9).
- Own-site figures are understated. Seven cited domains — aytimurtekstil.com, erenabiye.com, esergiyim.com.tr, matikabiyem.com, nihan.com.tr, tubaergin.com.tr and vogue.com.tr — look like a tracked entity’s own site but are not attributed to it, because the registry holds a different website.
- Qualifying positions on fewer than three models. Thirty qualifying entity-cut positions rest on one or two models. On the general cut: Sagaza (2 models, 66 mentions), Berneva (1 model, 50) and DCEY (2 models, 53). On the origin-constrained cut: Raisa Vanessa, Sagaza, Kayra, Berneva, Machka, Iconas Line and Exquise. On the origin-neutral cut: DCEY. The question-type cases are named in Chapter 6. Their visibility is model-specific, not panel-wide.
- Rank built on position. Beymen on the general cut, Raisa Vanessa on the origin-constrained and discovery cuts, and Faik Sönmez on the attribute cut rank high mainly because they are named early when named at all. That is the score working as designed, but it is a different kind of visibility from being named in a large share of all answers.
- Small score gaps are not meaningful. In the packed middle and lower part of each board — By Saygı (40.35) and Setrms (40.36), or Machka (35.10) and Kayra (34.92) — differences of a point or less should not be over-interpreted.
- Components can exceed 100. Because each component is scaled to the qualified leader and never clipped, 88 of 1,730 entity-cut rows (5.10%) exceed 100 on a component — all of them unqualified.
- Near-name pairs. Twelve pairs of similar names among entities named in at least three answers were flagged: Armut / Armut.com, Beymen Club / Beymen, DCEY / Dcey Bridal, Dekolte / EKOL, Farika / Fabrika, Kayra / Kayıra, Nihan / Nihan Peker, OGGUSTO / Gusto, Pazarama / Zara, Trendyolmilla / Trendyol, Tuğba / Bituğba Store and Vakkorama / Vakko. None shares a website on file, and each pair is tracked as two entities. The outcome is in the data; the reasoning behind each merge or non-merge is not.
- Trend rows on thin counts. Five rows in the two mover tables in Chapter 8 rest on fewer than 10 mentions in one edition — Nocturne ⚠, Kayra ⚠ and Armine ⚠ in July’s recounted answers, Yeşim Gelinlik ⚠ and Modazehrada ⚠ in October’s — and are marked ⚠ wherever they appear.
- System prompt and locale. Neither was recorded with this run.
- Single, point-in-time measurement. All responses were collected on 2 October 2026. Repeats within each question-model cell capture variation within a model, not drift from day to day.
- Models are probabilistic. The same question can produce different answers; repeats reduce this variability but do not remove it.
- Web search depends on momentary conditions. Grounded answers depend on what was indexed and served at the time of collection.
- Precision over recall. A brand or company is counted only where it is named in the answer, not where it appears only in a web address, and a shop named only as the place to buy is often not counted. Brands referred to indirectly are not counted either.
- Visibility is not quality. This report measures mention, position and breadth only. It does not assess whether a brand was described positively, whether a recommendation was accurate, or whether a dress is well made. A score is not an endorsement.
- Model versions date quickly. These findings are specific to the model versions in section 2.1, as of October 2026.
12. Appendix & Data
One-sentence takeaway: All 40 questions and all 393 tracked entities behind this report are listed below, and the underlying data is available on request.
12.1 All 40 questions
| # | Type | Scope | Question (Turkish) |
|---|---|---|---|
| 1 | Attribute | Origin-neutral | yazın terletmeyen abiye elbise |
| 2 | Use case | Origin-neutral | nikah için sade beyaz elbise nereden alınır |
| 3 | Attribute | Origin-neutral | göbek ve basen kapatan abiye nereden alınır |
| 4 | Attribute | Origin-constrained | türk malı büyük beden abiye önerisi |
| 5 | Use case | Origin-constrained | yerli markadan anne abiyesi arıyorum |
| 6 | Discovery | Origin-constrained | yerli mezuniyet elbisesi önerisi |
| 7 | Attribute | Origin-constrained | yerli tesettür abiye zayıf gösteren modeller |
| 8 | Attribute | Origin-neutral | saten abiye hangi marka iyi |
| 9 | Use case | Origin-neutral | hamile abiye nereden bulunur |
| 10 | Use case | Origin-constrained | türk malı kına elbisesi önerir misin |
| 11 | Use case | Origin-constrained | yerli markalardan mezuniyet elbisesi bakıyorum |
| 12 | Attribute | Origin-constrained | yerli saten abiye arıyorum |
| 13 | Use case | Origin-neutral | salon düğününde giymelik uzun abiye |
| 14 | Attribute | Origin-neutral | büyük beden abiye güzel duran markalar |
| 15 | Discovery | Origin-neutral | davet elbisesi nereden alınır |
| 16 | Use case | Origin-neutral | kına gecesi için kırmızı elbise önerisi |
| 17 | Use case | Origin-constrained | tesettürlü biri için yerli düğün abiyesi |
| 18 | Discovery | Origin-constrained | türk malı davet elbisesi önerir misin |
| 19 | Discovery | Origin-constrained | iyi bir yerli abiye markası var mı |
| 20 | Attribute | Origin-neutral | kısa boylular için abiye önerisi |
| 21 | Attribute | Origin-neutral | zayıf gösteren abiye elbise önerisi |
| 22 | Discovery | Origin-neutral | şık gece elbisesi hangi marka iyi |
| 23 | Discovery | Origin-constrained | davet elbisesi için yerli marka arıyorum |
| 24 | Discovery | Origin-constrained | Türk markalarından şık gece elbisesi |
| 25 | Use case | Origin-neutral | oğlumun mezuniyetinde giyeceğim şık elbise |
| 26 | Discovery | Origin-neutral | en iyi abiye markaları |
| 27 | Discovery | Origin-constrained | yerli tesettür abiye nereden alınır |
| 28 | Use case | Origin-constrained | yerli marka kısa mezuniyet elbisesi |
| 29 | Discovery | Origin-constrained | kaliteli yerli abiye nereden bakayım |
| 30 | Discovery | Origin-neutral | düğün misafiri için elbise nereden bakılır |
| 31 | Use case | Origin-constrained | Türk markalarından nikah sonrası yemek elbisesi |
| 32 | Use case | Origin-constrained | evde söz için sade yerli elbise |
| 33 | Attribute | Origin-constrained | iç göstermeyen astarlı yerli abiye nereden alınır |
| 34 | Attribute | Origin-constrained | kaliteli kumaşlı yerli davet elbisesi |
| 35 | Discovery | Origin-neutral | abiye elbise önerisi |
| 36 | Discovery | Origin-neutral | mezuniyet elbisesi markaları |
| 37 | Use case | Origin-neutral | arkadaşımın kır düğünü için elbise önerisi |
| 38 | Attribute | Origin-constrained | uygun fiyatlı yerli abiye markası |
| 39 | Discovery | Origin-neutral | tesettür abiye hangi siteden alınır |
| 40 | Attribute | Origin-neutral | uygun fiyatlı kaliteli abiye |
12.2 All tracked entities (393)
The table lists all 393 entities tracked in this study, ordered by general-cut mentions, and marks the 38 that qualify on the general cut. 65 entities qualify in at least one cut. Entities with no origin on file are shown as Unknown. Brand pages are not linked in this edition.
| Brand | Origin | Mentions | Rate | Models | Status |
|---|---|---|---|---|---|
| Trendyol | TR | 424 | 42.44% | 5 | Ranked |
| Carmen | TR | 384 | 38.44% | 5 | Ranked |
| ADL | TR | 260 | 26.03% | 5 | Ranked |
| The ZKS | TR | 258 | 25.83% | 5 | Ranked |
| AlfaBeta | TR | 242 | 24.22% | 5 | Ranked |
| Abiyefon | TR | 233 | 23.32% | 5 | Ranked |
| Sefamerve | TR | 213 | 21.32% | 5 | Ranked |
| Koton | TR | 196 | 19.62% | 5 | Ranked |
| Modanisa | TR | 188 | 18.82% | 5 | Ranked |
| Hepsiburada | TR | 178 | 17.82% | 4 | Ranked |
| Tarık Ediz | TR | 156 | 15.62% | 5 | Ranked |
| GIZIA | TR | 139 | 13.91% | 5 | Ranked |
| Kenzel | TR | 120 | 12.01% | 5 | Ranked |
| Cengiz Aktürk | TR | 111 | 11.11% | 4 | Ranked |
| Nocturne | TR | 109 | 10.91% | 4 | Ranked |
| İpekyol | TR | 109 | 10.91% | 4 | Ranked |
| Vakko | TR | 107 | 10.71% | 5 | Ranked |
| Roman | TR | 104 | 10.41% | 5 | Ranked |
| Setre | TR | 96 | 9.61% | 5 | Ranked |
| Rengin | TR | 94 | 9.41% | 5 | Ranked |
| Zara | Foreign | 83 | 8.31% | 5 | Ranked |
| Lafaba | TR | 77 | 7.71% | 5 | Ranked |
| Beymen | TR | 76 | 7.61% | 5 | Ranked |
| Kayra | TR | 74 | 7.41% | 3 | Ranked |
| Setrms | TR | 74 | 7.41% | 4 | Ranked |
| Armine | TR | 73 | 7.31% | 4 | Ranked |
| Machka | TR | 72 | 7.21% | 3 | Ranked |
| Oleg Cassini | Foreign | 72 | 7.21% | 3 | Ranked |
| LC Waikiki | TR | 67 | 6.71% | 4 | Ranked |
| Sagaza | TR | 66 | 6.61% | 2 | Ranked |
| EKOL | TR | 65 | 6.51% | 4 | Ranked |
| Mango | Foreign | 65 | 6.51% | 5 | Ranked |
| ModaSelvim | TR | 63 | 6.31% | 5 | Ranked |
| Iconas Line | TR | 61 | 6.11% | 4 | Ranked |
| Twist | TR | 61 | 6.11% | 5 | Ranked |
| DCEY | TR | 53 | 5.31% | 2 | Ranked |
| Berneva | TR | 50 | 5.01% | 1 | Ranked |
| By Saygı | TR | 50 | 5.01% | 4 | Ranked |
| Alvina | TR | 48 | 4.80% | 3 | Tracked |
| Eser Giyim | TR | 48 | 4.80% | 5 | Tracked |
| Trendyolmilla | TR | 46 | 4.60% | 4 | Tracked |
| H&M | Foreign | 44 | 4.40% | 5 | Tracked |
| Nihan | TR | 41 | 4.10% | 3 | Tracked |
| Network | TR | 39 | 3.90% | 4 | Tracked |
| Pronovias | Foreign | 39 | 3.90% | 5 | Tracked |
| BSL | TR | 38 | 3.80% | 5 | Tracked |
| Eyyo Atelier | TR | 37 | 3.70% | 5 | Tracked |
| Raisa Vanessa | TR | 37 | 3.70% | 2 | Tracked |
| Deniz Butik | TR | 36 | 3.60% | 3 | Tracked |
| Tuvanam | TR | 36 | 3.60% | 1 | Tracked |
| Akakçe | TR | 34 | 3.40% | 3 | Tracked |
| Lacrima | TR | 32 | 3.20% | 4 | Tracked |
| Ju Istanbul | TR | 31 | 3.10% | 1 | Tracked |
| Manuka | TR | 31 | 3.10% | 5 | Tracked |
| Touche Privé | TR | 31 | 3.10% | 1 | Tracked |
| Aker | TR | 29 | 2.90% | 4 | Tracked |
| Arya Moda | TR | 29 | 2.90% | 5 | Tracked |
| Exquise | TR | 28 | 2.80% | 1 | Tracked |
| Kıyafet Sepeti | TR | 27 | 2.70% | 4 | Tracked |
| Tuba Ergin | TR | 27 | 2.70% | 2 | Tracked |
| Tuğba | TR | 27 | 2.70% | 1 | Tracked |
| N11 | TR | 26 | 2.60% | 2 | Tracked |
| Faik Sönmez | TR | 25 | 2.50% | 5 | Tracked |
| Hülya Keser | TR | 25 | 2.50% | 3 | Tracked |
| Gigii's | TR | 24 | 2.40% | 1 | Tracked |
| Nurmelek | TR | 24 | 2.40% | 3 | Tracked |
| Özgür Masur | TR | 24 | 2.40% | 2 | Tracked |
| Boyner | TR | 23 | 2.30% | 5 | Tracked |
| LİLAS XXL | TR | 23 | 2.30% | 4 | Tracked |
| DeFacto | TR | 22 | 2.20% | 5 | Tracked |
| Minel Aşk | TR | 22 | 2.20% | 1 | Tracked |
| Şule Giyim | TR | 22 | 2.20% | 5 | Tracked |
| Amine Hüma | TR | 21 | 2.10% | 3 | Tracked |
| Coupe Official | TR | 21 | 2.10% | 1 | Tracked |
| Ola Abiye | TR | 21 | 2.10% | 2 | Tracked |
| PULLIMM | Unknown | 21 | 2.10% | 2 | Tracked |
| Tozlu | TR | 21 | 2.10% | 3 | Tracked |
| Nedret Taciroğlu | TR | 20 | 2.00% | 1 | Tracked |
| Forever New | Foreign | 19 | 1.90% | 4 | Tracked |
| Hüseyin Baştürk | TR | 19 | 1.90% | 3 | Tracked |
| Museum of Fine Clothing | TR | 19 | 1.90% | 1 | Tracked |
| Glorvia | TR | 18 | 1.80% | 3 | Tracked |
| My Best Friends | TR | 18 | 1.80% | 2 | Tracked |
| Seçil | TR | 18 | 1.80% | 4 | Tracked |
| So Lady | TR | 18 | 1.80% | 1 | Tracked |
| Ömür Inn | TR | 18 | 1.80% | 4 | Tracked |
| Ardanewline | TR | 17 | 1.70% | 3 | Tracked |
| Tuvid XXL | TR | 17 | 1.70% | 3 | Tracked |
| Dolap | TR | 16 | 1.60% | 2 | Tracked |
| Doridorca | TR | 16 | 1.60% | 4 | Tracked |
| House of Ogan | TR | 16 | 1.60% | 1 | Tracked |
| Pınar Şems | TR | 16 | 1.60% | 2 | Tracked |
| Sorbe | TR | 16 | 1.60% | 1 | Tracked |
| SpringStore | TR | 16 | 1.60% | 3 | Tracked |
| VznModa | TR | 16 | 1.60% | 3 | Tracked |
| Çiçeksepeti | TR | 16 | 1.60% | 3 | Tracked |
| Abiye Merkezi | TR | 15 | 1.50% | 3 | Tracked |
| Butik Nilyam | TR | 15 | 1.50% | 3 | Tracked |
| Farika | TR | 15 | 1.50% | 1 | Tracked |
| Gardrops | TR | 15 | 1.50% | 2 | Tracked |
| Burak Baran Fashion | TR | 14 | 1.40% | 3 | Tracked |
| Düğün.com | TR | 14 | 1.40% | 3 | Tracked |
| Patırtı | TR | 14 | 1.40% | 2 | Tracked |
| RUE Les Créateurs | TR | 14 | 1.40% | 2 | Tracked |
| Saygın | TR | 14 | 1.40% | 2 | Tracked |
| Öznur | Unknown | 14 | 1.40% | 2 | Tracked |
| Betty & Sam | TR | 13 | 1.30% | 4 | Tracked |
| Elben Moda | Unknown | 13 | 1.30% | 3 | Tracked |
| HOX'S | TR | 13 | 1.30% | 3 | Tracked |
| Miss Diamond | TR | 13 | 1.30% | 1 | Tracked |
| Şeref Vural | TR | 13 | 1.30% | 3 | Tracked |
| Accort | TR | 12 | 1.20% | 3 | Tracked |
| Cercis | TR | 12 | 1.20% | 3 | Tracked |
| Cimri | TR | 12 | 1.20% | 2 | Tracked |
| Kevser Sarıoğlu | TR | 12 | 1.20% | 1 | Tracked |
| Nazzan | TR | 12 | 1.20% | 2 | Tracked |
| Nurtuba | TR | 12 | 1.20% | 3 | Tracked |
| Pazarium | TR | 12 | 1.20% | 2 | Tracked |
| Pierre Cardin | Foreign | 12 | 1.20% | 2 | Tracked |
| Sarar | TR | 12 | 1.20% | 2 | Tracked |
| Beyyoglu | TR | 11 | 1.10% | 3 | Tracked |
| Dans | TR | 11 | 1.10% | 2 | Tracked |
| Gusto | TR | 11 | 1.10% | 5 | Tracked |
| Nihan Peker | TR | 11 | 1.10% | 2 | Tracked |
| Olcay | TR | 11 | 1.10% | 4 | Tracked |
| ESSEDRESS | Unknown | 10 | 1.00% | 2 | Tracked |
| Elbise Kirala | TR | 10 | 1.00% | 4 | Tracked |
| Gamze Özkul | TR | 10 | 1.00% | 1 | Tracked |
| Mayabella | Unknown | 10 | 1.00% | 1 | Tracked |
| Pionto | TR | 10 | 1.00% | 1 | Tracked |
| Serica Moda | TR | 10 | 1.00% | 3 | Tracked |
| Tofisa | TR | 10 | 1.00% | 2 | Tracked |
| Zühre | TR | 10 | 1.00% | 3 | Tracked |
| Al-Tatari | TR | 9 | 0.90% | 4 | Tracked |
| Alia | TR | 9 | 0.90% | 2 | Tracked |
| Aysa Tekin | TR | 9 | 0.90% | 3 | Tracked |
| Aysira | TR | 9 | 0.90% | 2 | Tracked |
| Disentis Modest | TR | 9 | 0.90% | 1 | Tracked |
| Elie Saab | Foreign | 9 | 0.90% | 3 | Tracked |
| Ilmio | TR | 9 | 0.90% | 2 | Tracked |
| Neva Style | TR | 9 | 0.90% | 2 | Tracked |
| Perspective | TR | 9 | 0.90% | 1 | Tracked |
| PodyumPlus | TR | 9 | 0.90% | 3 | Tracked |
| Sheeny | Unknown | 9 | 0.90% | 3 | Tracked |
| Tarz Kadın | Unknown | 9 | 0.90% | 3 | Tracked |
| Venn | Unknown | 9 | 0.90% | 1 | Tracked |
| Yeşim Gelinlik | TR | 9 | 0.90% | 3 | Tracked |
| Zagrep | Unknown | 9 | 0.90% | 4 | Tracked |
| Beymen Club | TR | 8 | 0.80% | 2 | Tracked |
| Beyza Giyim | TR | 8 | 0.80% | 1 | Tracked |
| Beyza Online | TR | 8 | 0.80% | 1 | Tracked |
| Cengiz Abazoğlu | Unknown | 8 | 0.80% | 1 | Tracked |
| ESRAHELVACI | TR | 8 | 0.80% | 2 | Tracked |
| Gökçe Moda Evi | TR | 8 | 0.80% | 2 | Tracked |
| Hürrem Aldinç | TR | 8 | 0.80% | 2 | Tracked |
| Jovani | Foreign | 8 | 0.80% | 2 | Tracked |
| Leyuze Butik | TR | 8 | 0.80% | 2 | Tracked |
| MK Moda Tasarım | TR | 8 | 0.80% | 2 | Tracked |
| Nefeli Store | TR | 8 | 0.80% | 2 | Tracked |
| Odrica | TR | 8 | 0.80% | 2 | Tracked |
| Rmg | TR | 8 | 0.80% | 3 | Tracked |
| Stradivarius | Foreign | 8 | 0.80% | 3 | Tracked |
| Didem Soydan | Unknown | 7 | 0.70% | 1 | Tracked |
| DreamON | TR | 7 | 0.70% | 2 | Tracked |
| Dressland | TR | 7 | 0.70% | 1 | Tracked |
| E-Tesettür | Unknown | 7 | 0.70% | 1 | Tracked |
| GILLA'S QUEEN | Unknown | 7 | 0.70% | 1 | Tracked |
| Love My Body | TR | 7 | 0.70% | 2 | Tracked |
| Modacadiri | TR | 7 | 0.70% | 2 | Tracked |
| Modazehrada | TR | 7 | 0.70% | 3 | Tracked |
| Nika Fashion | TR | 7 | 0.70% | 2 | Tracked |
| Qupra | TR | 7 | 0.70% | 1 | Tracked |
| Raşit Bağzıbağlı | TR | 7 | 0.70% | 1 | Tracked |
| Rentony | TR | 7 | 0.70% | 2 | Tracked |
| Tesettür Island | TR | 7 | 0.70% | 3 | Tracked |
| Valentino | Foreign | 7 | 0.70% | 3 | Tracked |
| Yusem Tesettür | TR | 7 | 0.70% | 2 | Tracked |
| Zeynep Tosun | TR | 7 | 0.70% | 1 | Tracked |
| Aytimur Tekstil | Unknown | 6 | 0.60% | 3 | Tracked |
| Ceylan Otantik | TR | 6 | 0.60% | 1 | Tracked |
| Davet Çok Elbisem Yok | TR | 6 | 0.60% | 2 | Tracked |
| Davetten Davete | TR | 6 | 0.60% | 1 | Tracked |
| Dekolte | TR | 6 | 0.60% | 2 | Tracked |
| Dilvin | TR | 6 | 0.60% | 1 | Tracked |
| Fabrika | TR | 6 | 0.60% | 2 | Tracked |
| Giyze | TR | 6 | 0.60% | 2 | Tracked |
| Merci MS Fancy | TR | 6 | 0.60% | 2 | Tracked |
| Milagron | TR | 6 | 0.60% | 1 | Tracked |
| Mina Lidya | TR | 6 | 0.60% | 2 | Tracked |
| Modavina | TR | 6 | 0.60% | 2 | Tracked |
| OXXO | TR | 6 | 0.60% | 1 | Tracked |
| Tiara | TR | 6 | 0.60% | 2 | Tracked |
| Wom Boutique | TR | 6 | 0.60% | 1 | Tracked |
| Şans | TR | 6 | 0.60% | 2 | Tracked |
| A46 | Unknown | 5 | 0.50% | 1 | Tracked |
| ALFUDO | TR | 5 | 0.50% | 2 | Tracked |
| Abiyenial | TR | 5 | 0.50% | 2 | Tracked |
| Alisse Nuera | TR | 5 | 0.50% | 3 | Tracked |
| Babil Butik | TR | 5 | 0.50% | 1 | Tracked |
| Berissima Bride | TR | 5 | 0.50% | 2 | Tracked |
| Bituğba Store | TR | 5 | 0.50% | 2 | Tracked |
| Busa | TR | 5 | 0.50% | 2 | Tracked |
| CML Collection | Unknown | 5 | 0.50% | 2 | Tracked |
| Civil | TR | 5 | 0.50% | 1 | Tracked |
| Desigual | Foreign | 5 | 0.50% | 1 | Tracked |
| Dilek Hanif | TR | 5 | 0.50% | 1 | Tracked |
| Doucée | TR | 5 | 0.50% | 1 | Tracked |
| GiyzaMila | TR | 5 | 0.50% | 1 | Tracked |
| Gizem Butik | TR | 5 | 0.50% | 1 | Tracked |
| Gönül Kolat | TR | 5 | 0.50% | 1 | Tracked |
| Just Cavalli | Foreign | 5 | 0.50% | 1 | Tracked |
| Karl Lagerfeld | Foreign | 5 | 0.50% | 1 | Tracked |
| LACASE | Unknown | 5 | 0.50% | 2 | Tracked |
| Laranor | TR | 5 | 0.50% | 1 | Tracked |
| Massimo Dutti | Foreign | 5 | 0.50% | 2 | Tracked |
| Melek Araz | TR | 5 | 0.50% | 1 | Tracked |
| Meliha Tarhanacı | TR | 5 | 0.50% | 3 | Tracked |
| Nur Karaata | TR | 5 | 0.50% | 2 | Tracked |
| OGGUSTO | TR | 5 | 0.50% | 2 | Tracked |
| Puane | TR | 5 | 0.50% | 2 | Tracked |
| Seamoda | TR | 5 | 0.50% | 3 | Tracked |
| Self-Portrait | Foreign | 5 | 0.50% | 1 | Tracked |
| Sherri Hill | Foreign | 5 | 0.50% | 1 | Tracked |
| Societta Abiye | TR | 5 | 0.50% | 2 | Tracked |
| Spazio | TR | 5 | 0.50% | 2 | Tracked |
| United Colors of Benetton | Foreign | 5 | 0.50% | 1 | Tracked |
| UnoMoi | TR | 5 | 0.50% | 1 | Tracked |
| Zuhair Murad | Foreign | 5 | 0.50% | 1 | Tracked |
| AYŞEYAMAÇLI | Unknown | 4 | 0.40% | 1 | Tracked |
| Abiyecim | TR | 4 | 0.40% | 3 | Tracked |
| Akay Gelinlik | TR | 4 | 0.40% | 1 | Tracked |
| Ankyra | TR | 4 | 0.40% | 2 | Tracked |
| Ayşegül Kalelio | TR | 4 | 0.40% | 2 | Tracked |
| Azra Design | TR | 4 | 0.40% | 1 | Tracked |
| Brandroom | TR | 4 | 0.40% | 1 | Tracked |
| Ceylen | TR | 4 | 0.40% | 1 | Tracked |
| Chanel | Foreign | 4 | 0.40% | 2 | Tracked |
| Efelif | TR | 4 | 0.40% | 2 | Tracked |
| Eren Abiye | TR | 4 | 0.40% | 1 | Tracked |
| InStyle | Unknown | 4 | 0.40% | 2 | Tracked |
| Jeyan Gedik | TR | 4 | 0.40% | 1 | Tracked |
| Kaftan Collection | TR | 4 | 0.40% | 1 | Tracked |
| Kayıra | Unknown | 4 | 0.40% | 1 | Tracked |
| Mac Duggal | Foreign | 4 | 0.40% | 1 | Tracked |
| Modarana | TR | 4 | 0.40% | 2 | Tracked |
| Muse for All | TR | 4 | 0.40% | 1 | Tracked |
| Needle & Thread | Foreign | 4 | 0.40% | 1 | Tracked |
| Nursena Moda | TR | 4 | 0.40% | 2 | Tracked |
| Ottomanwear | TR | 4 | 0.40% | 1 | Tracked |
| Reformation | Foreign | 4 | 0.40% | 3 | Tracked |
| Rent Code | Unknown | 4 | 0.40% | 1 | Tracked |
| Rent The Look | Unknown | 4 | 0.40% | 1 | Tracked |
| Rivus | TR | 4 | 0.40% | 1 | Tracked |
| Seninmoda | TR | 4 | 0.40% | 1 | Tracked |
| Sima Couture | Foreign | 4 | 0.40% | 1 | Tracked |
| Solace London | Foreign | 4 | 0.40% | 1 | Tracked |
| Sunfere | Foreign | 4 | 0.40% | 2 | Tracked |
| Terani | Foreign | 4 | 0.40% | 1 | Tracked |
| Tesettür Home | TR | 4 | 0.40% | 1 | Tracked |
| Vizyon La Mira | TR | 4 | 0.40% | 1 | Tracked |
| Vogue | Foreign | 4 | 0.40% | 1 | Tracked |
| Zimmermann | Foreign | 4 | 0.40% | 1 | Tracked |
| Zishan | TR | 4 | 0.40% | 2 | Tracked |
| eHamileGiyim | TR | 4 | 0.40% | 1 | Tracked |
| İmajButik | TR | 4 | 0.40% | 1 | Tracked |
| İpek Kıramer | TR | 4 | 0.40% | 2 | Tracked |
| Şirin Butik | TR | 4 | 0.40% | 3 | Tracked |
| ASOS | Foreign | 3 | 0.30% | 2 | Tracked |
| Afife | Unknown | 3 | 0.30% | 2 | Tracked |
| Angelino | TR | 3 | 0.30% | 1 | Tracked |
| Armani Exchange | Foreign | 3 | 0.30% | 1 | Tracked |
| Armut | TR | 3 | 0.30% | 2 | Tracked |
| Armut.com | Unknown | 3 | 0.30% | 1 | Tracked |
| Arıkan | TR | 3 | 0.30% | 1 | Tracked |
| Azazie | Foreign | 3 | 0.30% | 2 | Tracked |
| Aşkın Göksu | TR | 3 | 0.30% | 2 | Tracked |
| BDM Exclusive | TR | 3 | 0.30% | 1 | Tracked |
| Burcum By B.C.C. | Unknown | 3 | 0.30% | 2 | Tracked |
| Bürün | Unknown | 3 | 0.30% | 1 | Tracked |
| Büyük Bedeniz | TR | 3 | 0.30% | 1 | Tracked |
| Chilia Design | TR | 3 | 0.30% | 1 | Tracked |
| Dcey Bridal | TR | 3 | 0.30% | 1 | Tracked |
| Devr-i Tesettür | TR | 3 | 0.30% | 2 | Tracked |
| Divina Wedding | TR | 3 | 0.30% | 1 | Tracked |
| Emsale | TR | 3 | 0.30% | 2 | Tracked |
| Esma Karadağ | TR | 3 | 0.30% | 3 | Tracked |
| Gamze Polat | Unknown | 3 | 0.30% | 1 | Tracked |
| Gebe | TR | 3 | 0.30% | 1 | Tracked |
| Glr & Gym | Unknown | 3 | 0.30% | 2 | Tracked |
| HESNA | TR | 3 | 0.30% | 1 | Tracked |
| Işşıl | TR | 3 | 0.30% | 1 | Tracked |
| Jacquemus | Foreign | 3 | 0.30% | 1 | Tracked |
| Kameya | TR | 3 | 0.30% | 1 | Tracked |
| Keikei | TR | 3 | 0.30% | 2 | Tracked |
| Kırmızı Butik | Unknown | 3 | 0.30% | 1 | Tracked |
| Lamia Giyim | TR | 3 | 0.30% | 3 | Tracked |
| Lirenza Abiye | TR | 3 | 0.30% | 1 | Tracked |
| Love on Friday | TR | 3 | 0.30% | 1 | Tracked |
| Lulus | Foreign | 3 | 0.30% | 2 | Tracked |
| Matik Abiyem | Unknown | 3 | 0.30% | 1 | Tracked |
| MegaButik | TR | 3 | 0.30% | 2 | Tracked |
| Merhaba Boutique | TR | 3 | 0.30% | 2 | Tracked |
| Mevra | TR | 3 | 0.30% | 2 | Tracked |
| Moda Berray | TR | 3 | 0.30% | 1 | Tracked |
| Modafeminen | Unknown | 3 | 0.30% | 2 | Tracked |
| Modapinhan | TR | 3 | 0.30% | 1 | Tracked |
| Modatalika | TR | 3 | 0.30% | 1 | Tracked |
| Mossta | TR | 3 | 0.30% | 1 | Tracked |
| Nefise | TR | 3 | 0.30% | 1 | Tracked |
| Next | Foreign | 3 | 0.30% | 2 | Tracked |
| Patrizia Pepe | Foreign | 3 | 0.30% | 1 | Tracked |
| Pazarama | TR | 3 | 0.30% | 1 | Tracked |
| Pinko | Foreign | 3 | 0.30% | 1 | Tracked |
| Refka Abiye | TR | 3 | 0.30% | 2 | Tracked |
| SUBU TASARIM | Unknown | 3 | 0.30% | 1 | Tracked |
| Sedef | Unknown | 3 | 0.30% | 1 | Tracked |
| Semala Butik | TR | 3 | 0.30% | 1 | Tracked |
| Setr-i Nur | TR | 3 | 0.30% | 1 | Tracked |
| Tommy Hilfiger | Foreign | 3 | 0.30% | 1 | Tracked |
| Trend Tesettür | TR | 3 | 0.30% | 2 | Tracked |
| Vakkorama | TR | 3 | 0.30% | 2 | Tracked |
| Vipiri | Unknown | 3 | 0.30% | 1 | Tracked |
| Zoezoe | Unknown | 3 | 0.30% | 1 | Tracked |
| Züleyha Kuru | TR | 3 | 0.30% | 1 | Tracked |
| lovebox | TR | 3 | 0.30% | 1 | Tracked |
| vuvutasarım | Unknown | 3 | 0.30% | 1 | Tracked |
| ALLDAY | TR | 2 | 0.20% | 1 | Tracked |
| Anthropologie | Foreign | 2 | 0.20% | 2 | Tracked |
| Arzu Kaprol | TR | 2 | 0.20% | 2 | Tracked |
| BAQA | TR | 2 | 0.20% | 1 | Tracked |
| Büyük Moda | TR | 2 | 0.20% | 2 | Tracked |
| David's Bridal | Foreign | 2 | 0.20% | 1 | Tracked |
| Denokids | TR | 2 | 0.20% | 1 | Tracked |
| DressIstanbul | TR | 2 | 0.20% | 1 | Tracked |
| Forever 21 | Foreign | 2 | 0.20% | 1 | Tracked |
| Givenchy SA | Foreign | 2 | 0.20% | 1 | Tracked |
| Gül Hürgel | TR | 2 | 0.20% | 2 | Tracked |
| HİRAMİR | TR | 2 | 0.20% | 1 | Tracked |
| LOEWE | Foreign | 2 | 0.20% | 1 | Tracked |
| Lenta Moda | TR | 2 | 0.20% | 2 | Tracked |
| Mavi | TR | 2 | 0.20% | 1 | Tracked |
| Miraggio | TR | 2 | 0.20% | 1 | Tracked |
| Modailgi | TR | 2 | 0.20% | 2 | Tracked |
| Morhipo | TR | 2 | 0.20% | 1 | Tracked |
| SIRRI | TR | 2 | 0.20% | 1 | Tracked |
| Sateen | TR | 2 | 0.20% | 1 | Tracked |
| Suud Collection | TR | 2 | 0.20% | 1 | Tracked |
| Tuay Karaca | TR | 2 | 0.20% | 2 | Tracked |
| Abercrombie & Fitch | Foreign | 1 | 0.10% | 1 | Tracked |
| AbiyeElbisecim | Unknown | 1 | 0.10% | 1 | Tracked |
| Alexander McQueen | Foreign | 1 | 0.10% | 1 | Tracked |
| Alexandre Vauthier | Foreign | 1 | 0.10% | 1 | Tracked |
| Balmain Türkiye | Foreign | 1 | 0.10% | 1 | Tracked |
| Bershka Türkiye | Foreign | 1 | 0.10% | 1 | Tracked |
| Beyliss | TR | 1 | 0.10% | 1 | Tracked |
| Davet Elegance | TR | 1 | 0.10% | 1 | Tracked |
| Dior | Foreign | 1 | 0.10% | 1 | Tracked |
| Dolce&Gabbana | Foreign | 1 | 0.10% | 1 | Tracked |
| ESOTTE | TR | 1 | 0.10% | 1 | Tracked |
| ETRO | Foreign | 1 | 0.10% | 1 | Tracked |
| Enderun | Unknown | 1 | 0.10% | 1 | Tracked |
| Etsy | Foreign | 1 | 0.10% | 1 | Tracked |
| GUESS | Foreign | 1 | 0.10% | 1 | Tracked |
| Ghisa | TR | 1 | 0.10% | 1 | Tracked |
| Gülhan Moda | TR | 1 | 0.10% | 1 | Tracked |
| Hanezza | TR | 1 | 0.10% | 1 | Tracked |
| Haza Moda | TR | 1 | 0.10% | 1 | Tracked |
| House of CB | Foreign | 1 | 0.10% | 1 | Tracked |
| Işıltı Abiye | TR | 1 | 0.10% | 1 | Tracked |
| Jeep | Foreign | 1 | 0.10% | 1 | Tracked |
| KIKIRIKI | TR | 1 | 0.10% | 1 | Tracked |
| Kadriye Baştürk | TR | 1 | 0.10% | 1 | Tracked |
| Lady Special | TR | 1 | 0.10% | 1 | Tracked |
| Le Mabelle | TR | 1 | 0.10% | 1 | Tracked |
| Lee Cooper | Foreign | 1 | 0.10% | 1 | Tracked |
| Los Banditos | TR | 1 | 0.10% | 1 | Tracked |
| Macy's | Foreign | 1 | 0.10% | 1 | Tracked |
| Moda Babil | TR | 1 | 0.10% | 1 | Tracked |
| Modasena | TR | 1 | 0.10% | 1 | Tracked |
| Mudo | TR | 1 | 0.10% | 1 | Tracked |
| My Rental Dress | TR | 1 | 0.10% | 1 | Tracked |
| NAZRA | Unknown | 1 | 0.10% | 1 | Tracked |
| Nordstrom | Foreign | 1 | 0.10% | 1 | Tracked |
| Panço | TR | 1 | 0.10% | 1 | Tracked |
| SAMARA | Unknown | 1 | 0.10% | 1 | Tracked |
| Siren Ertan | TR | 1 | 0.10% | 1 | Tracked |
| Styling Park | TR | 1 | 0.10% | 1 | Tracked |
| TOM FORD | Foreign | 1 | 0.10% | 1 | Tracked |
| Versace | Foreign | 1 | 0.10% | 1 | Tracked |
| Victoria Beckham | Foreign | 1 | 0.10% | 1 | Tracked |
| Zülays | TR | 1 | 0.10% | 1 | Tracked |
| klodsy | TR | 1 | 0.10% | 1 | Tracked |
| letgo | TR | 1 | 0.10% | 1 | Tracked |
12.3 Data availability
The data behind this report is available on request: per-entity metrics across all six cuts, the qualified leaderboards and their brand-only variants, the entity registry with origins, the per-model breakdowns, the origin-share tables on both bases, the source attribution tables, the raw-layer trend file, and the published scale registry.
All figures reflect a single collection day (2 October 2026) and are subject to the limitations in Chapter 11. This is the second edition; the study is repeated quarterly.
Methodology note
Frequently asked questions
Across all 40 questions, the marketplace Trendyol leads with a 100.00 score, named in 42.44% of 999 responses and by all five assistants. Carmen (93.77) is second, named in 38.44% of responses, and The ZKS, ADL and Abiyefon complete the top five. Carmen leads the 20 questions that ask for Turkish brands.
Mostly Turkish. On the 20 questions that ask for Turkish brands, domestic entities take 100.00% of qualifying mentions and foreign entities 0.00%. On the 20 origin-neutral questions they take 83.21% and 16.79%, and five foreign brands qualify: Mango, H&M, Zara, Oleg Cassini and Pronovias. Origin is on file for every qualifying mention on both.
393 entities were tracked, including retailers and marketplaces. 38 cleared the 5% mention threshold on the full question set — 35 domestic and 3 foreign — with an average score of 52.64, and 65 qualify in at least one cut.
It combines how often an entity is named (45%), how early it appears in the answer (30%), and how many of the five models name it (25%). Each part is scaled against the leading qualified entity within that cut, so scores are never compared across cuts; across editions, only the general board's change since July is shown, against July re-scored the same way. Only entities named in at least 5% of a cut's responses are ranked.
On the leaders, largely: Carmen is in every model's top three, and Trendyol, Carmen, ADL and AlfaBeta are in every model's top ten. On length, no: Gemini 3.6 Flash names 12.38 entities per answer and Claude Haiku 4.5 names 4.44. Some ranks lean on one model: ChatGPT 5.6 Luna Pro supplies all 50 of Berneva's mentions, and Grok 4.3 supplies 111 of Hepsiburada's 178.
The 40 questions are the same, but two of the five models were replaced, so changes show direction more reliably than size. Measured by raw mention rate against July's answers counted the way this edition counts, The ZKS gained 19.01 points and Iconas Line lost 8.54. Origin shares are not compared with July's, because the origin rule changed.
No. The score measures only how often and how prominently a brand is named across the assistants. It does not assess quality, fit, price, accuracy or sentiment, so a score is a measure of visibility, not an endorsement.
About Herm.io & disclosure
Herm.io is a consumer behaviour and marketing data company. We study how people discover and choose brands to help businesses reach the right customers. This report forms part of our public research and is conducted quarterly.
Disclosure & neutrality
The rankings and metrics in this report are entirely impartial and are based on the methodology described within the report. No brand can pay or provide sponsorship to be included in the report, improve its ranking or influence how it is described in the content.
A brand’s score in the report is not an endorsement or assessment of quality; it is solely a measure of its current visibility within AI models. No service or product offered by Herm.io is sold for the purpose of directly changing the results of this report or guaranteeing a ranking. To preserve objectivity, all references to Herm.io’s own domain are removed from the source data before analysis. This ensures that the company does not appear in, measure itself through or benefit from its own study.
Brands seeking to understand their position in the data can schedule a consultation for an impartial assessment of the findings. The consultation is advisory and free of charge.
Mert Can Elkaya
Contributor
I'm a product builder working at the intersection of product, fintech, and growth. From martech and venture capital to leading product at a proptech platform and co-founding a fintech startup, I help teams—and shoppers—make smarter, more confident decisions.
More from Mert