# Synthetic Consumer Lab (SCL)

> Synthetic Consumer Lab (SCL) is an İstanbul-based research platform. Companies test prices, products, packaging, ads and messages on synthetic consumers: AI agents, each with a realistic demographic and socioeconomic profile built from statistical distributions and past consumer research, with ready-made samples for Türkiye. In consumer research these agents are also called synthetic respondents or digital twins of consumers, and the answers they give are synthetic data. Supported methods: willingness to pay (WTP), conjoint, MaxDiff, multiple-choice surveys, focus groups, ad tests, message tests and A/B tests. Results are model outputs that indicate direction and magnitude; SCL compares its models with real consumer data and recommends validating high-stakes decisions with real consumers.

The site is bilingual: English under /en, Turkish at the root. Contact: info@synthetic-consumers.com. A single-file version of the key pages is at https://synthetic-consumers.com/llms-full.txt

## Guides: definitions (English)

- [Digital twins of consumers: what they are](https://synthetic-consumers.com/en/digital-twins): A digital twin of a consumer is an AI model of a real consumer profile. How digital twins are used in market research, how they differ from personas, and their limits.
- [What is synthetic data? Its use in consumer research](https://synthetic-consumers.com/en/synthetic-data): Synthetic data is artificial data that keeps the statistical properties of real data. How it is generated for consumer research, when to trust it, and where its limits are.
- [What is a synthetic survey? Synthetic respondents](https://synthetic-consumers.com/en/synthetic-surveys): A synthetic survey puts your questions to synthetic respondents instead of real people. How it works, which questions it suits, and how its accuracy is tested.

## Product (English)

- [Synthetic consumer research: synthetic surveys and data](https://synthetic-consumers.com/en): Test product, pricing and advertising ideas with synthetic consumers. Run synthetic surveys with digital twins of consumers and get results in minutes.
- [About us: the synthetic consumer research team](https://synthetic-consumers.com/en/about): Synthetic Consumer Lab is an İstanbul-based team that runs research with digital twins of consumers instead of fieldwork. Meet the team and our approach.
- [Demo: ask a thousand synthetic consumers a question](https://synthetic-consumers.com/en/demo): Ask one question and let a thousand synthetic consumers answer. Try a scaled-down version of the SCL panel.
- [Early access](https://synthetic-consumers.com/en/early-access): SCL early access: share your research question, design your first study with synthetic consumers together with us, and be among the first to try the demo.
- [How it works: synthetic consumer research](https://synthetic-consumers.com/en/how-it-works): From question to audience, simulation to report: the four steps of preparing a synthetic survey or consumer study with SCL, and how to read the results.
- [Pricing: synthetic consumer research](https://synthetic-consumers.com/en/pricing): Estimate the cost of your synthetic consumer research. Pricing per Person-Question, a free starter account and enterprise packages.

## Published studies (English)

- [Studies: research with synthetic consumers](https://synthetic-consumers.com/en/studies): Open studies run with synthetic consumers: price sensitivity in snacks, and how social proof changes willingness to pay for headphones.
- [What do we really trust when buying a product?](https://synthetic-consumers.com/en/studies/kulaklik-wtp-karsilastirma): The same headphones under four social-proof scenarios: we measured with synthetic consumers how ratings and reviews change willingness to pay (WTP).
- [No review information — WTP study](https://synthetic-consumers.com/en/studies/kulaklik-wtp-karsilastirma/degerlendirme-yok): No review information — WTP study: price acceptance rates, the demand curve and willingness to pay for the headphones in this scenario. A study run with synthetic consumers.
- [Negative review — WTP study](https://synthetic-consumers.com/en/studies/kulaklik-wtp-karsilastirma/olumsuz-yorum): Negative review — WTP study: price acceptance rates, the demand curve and willingness to pay for the headphones in this scenario. A study run with synthetic consumers.
- [High rating + few reviews — WTP study](https://synthetic-consumers.com/en/studies/kulaklik-wtp-karsilastirma/yuksek-puan-az-degerlendirme): High rating + few reviews — WTP study: price acceptance rates, the demand curve and willingness to pay for the headphones in this scenario. A study run with synthetic consumers.
- [High rating + many reviews — WTP study](https://synthetic-consumers.com/en/studies/kulaklik-wtp-karsilastirma/yuksek-puan-cok-degerlendirme): High rating + many reviews — WTP study: price acceptance rates, the demand curve and willingness to pay for the headphones in this scenario. A study run with synthetic consumers.
- [Price sensitivity in snack products](https://synthetic-consumers.com/en/studies/snack-price-sensitivity): Price acceptance, sharp drops in demand and willingness to pay by segment for two snack products: a price sensitivity study run with synthetic consumers.
- [Chocolate-coated Hazelnut Wafer — Price sensitivity](https://synthetic-consumers.com/en/studies/snack-price-sensitivity/chocolate-coated-hazelnut-wafer): Chocolate-coated Hazelnut Wafer: expected demand across the tested price range, the best tested price and willingness to pay by segment. A study run with synthetic consumers.
- [Multipack Cocoa Sandwich Cake — Price sensitivity](https://synthetic-consumers.com/en/studies/snack-price-sensitivity/cocoa-sandwich-cake): Multipack Cocoa Sandwich Cake: expected demand across the tested price range, the best tested price and willingness to pay by segment. A study run with synthetic consumers.

## Research library: summaries of academic papers on LLM-simulated respondents (English)

- [Articles: simulating consumers with LLMs](https://synthetic-consumers.com/en/articles): Summaries and key findings of academic papers on synthetic respondents and simulating consumers with large language models.
- [Analyzing Brand Perception in LLMs](https://synthetic-consumers.com/en/articles/analyzing-brand-perception-in-llms): Evaluating how biased LLMs are towards brands and how consistent their answers are.
- [Beyond Demographics: Aligning Role-playing LLM-based Agents Using Human Belief Networks](https://synthetic-consumers.com/en/articles/beyond-demographics-aligning-role-playing-llm-based-agents-using-human-belief-networks): How accurately LLM agents imitate human opinions when given demographic information and belief networks.
- [Can LLMs Capture Human Preferences?](https://synthetic-consumers.com/en/articles/can-llms-capture-human-preferences): The ability of LLMs to imitate human preferences and the effects of language and chain-of-thought methods.
- [Comparing Traditional and LLM-based Search for Consumer Choice: A Randomized Experiment](https://synthetic-consumers.com/en/articles/comparing-traditional-and-llm-based-search-for-consumer-choice-a-randomized-experiment): How LLM-based search affects efficiency, accuracy and user perception compared with traditional search, and the role of confidence cues.
- [Do LLM Agents Exhibit Social Behavior?](https://synthetic-consumers.com/en/articles/do-llm-agents-exhibit-social-behavior): How LLMs behave in social interactions: the role of group identity, reciprocity and chain-of-thought reasoning.
- [Enhancing LLM Decision-Making with Factor Profiles and Analogical Reasoning (DEFINE)](https://synthetic-consumers.com/en/articles/enhancing-llm-decision-making-with-factor-profiles-and-analogical-reasoning-define): Rational decision-making by LLMs in financial scenarios under uncertainty: improving decision accuracy and transparency through factor profiles and historical analogy.
- [Large Language Models as Simulated Economic Agents: What Can We Learn from Homo Silicus?](https://synthetic-consumers.com/en/articles/large-language-models-as-simulated-economic-agents-what-can-we-learn-from-homo-silicus): The capacity of LLMs to imitate human agents in behavioural economics experiments: the effect of endowing models with different attributes and the reproducibility of results.
- [Large Language Models as Subpopulation Representative Models: A Review](https://synthetic-consumers.com/en/articles/large-language-models-as-subpopulation-representative-models-a-review): Using LLMs as subpopulation representative models (SRMs): potential for measuring public opinion and a risk–benefit assessment.
- [Large Language Models Empowered Agent-Based Modeling and Simulation: A Survey and Perspectives](https://synthetic-consumers.com/en/articles/large-language-models-empowered-agent-based-modeling-and-simulation-a-survey-and-perspectives): The role of LLMs in agent-based modelling and simulation (ABMS): transforming the simulation paradigm through human-like intelligence and behaviour.
- [LLM-Based Doppelgänger Models: Leveraging Synthetic Data for Human-Like Responses](https://synthetic-consumers.com/en/articles/llm-based-doppelganger-models-leveraging-synthetic-data-for-human-like-responses-in-survey-simulations): Learning and imitating individual opinions with an LLM-based Doppelgänger model: accuracy, heterogeneity, and single-device feasibility.
- [LLMs and Generative Agent-Based Models for Complex Systems](https://synthetic-consumers.com/en/articles/llms-and-generative-agent-based-models-for-complex-systems): The capacity of LLM agents to imitate human behaviour across social systems, and the modelling of complex social dynamics with Generative Agent-Based Models (GABMs).
- [LLMs’ Ways of Seeing User Personas](https://synthetic-consumers.com/en/articles/llms-ways-of-seeing-user-personas): LLMs’ ability to perceive and interpret user personas: the impact of cultural context and the capacity to reconstruct demographic profiles.
- [MatrAIx: Simulating the World with 8.3 Billion Persona Agents](https://synthetic-consumers.com/en/articles/matraix-simulating-the-world-with-8-3-billion-persona-agents): A population-scale simulated-user evaluation infrastructure that tests AI systems and digital products with 8.3 billion persona records and 1,010 reusable tasks.
- [Reduce Disparity Between LLMs and Humans: Optimal LLM Sample Calibration](https://synthetic-consumers.com/en/articles/reduce-disparity-between-llms-and-humans-optimal-llm-sample-calibration): Calibration of LLMs across demographic groups: human alignment and transferability using Human Mimicry Calibration (HMC).
- [Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization](https://synthetic-consumers.com/en/articles/two-tales-of-persona-in-llms-a-survey-of-role-playing-and-personalization): Persona usage in LLMs: the distinction between Role-Playing and Personalization, application domains, and personality assessment methods.
- [Using Large Language Models to Simulate Multiple Humans and Replicate Human Subject Studies](https://synthetic-consumers.com/en/articles/using-large-language-models-to-simulate-multiple-humans-and-replicate-human-subject-studies): The capacity of LLMs to simulate human behaviour across classic experiments and diverse demographic profiles: model scale, demographic variation, and hyper-accuracy effects.

## Rehberler: tanımlar (Türkçe)

- [Dijital ikiz nedir? Tüketici dijital ikizleri](https://synthetic-consumers.com/dijital-ikiz): Tüketici dijital ikizi, gerçek bir tüketici profilinin yapay zekâ ile modellenmiş karşılığıdır. Dijital ikizler pazar araştırmasında nasıl kullanılır, sınırları nelerdir?
- [Sentetik anket nedir? Sentetik katılımcılarla araştırma](https://synthetic-consumers.com/sentetik-anket): Sentetik anket, soruların gerçek kişiler yerine sentetik katılımcılara sorulduğu araştırmadır. Nasıl çalışır, hangi sorulara uygundur, doğruluğu nasıl sınanır?
- [Sentetik veri nedir? Tüketici araştırmasında kullanımı](https://synthetic-consumers.com/sentetik-veri): Sentetik veri, gerçek verinin istatistiksel özelliklerini taşıyan yapay veridir. Tüketici araştırmasında nasıl üretilir, ne zaman güvenilir, sınırları nelerdir?

## Ürün (Türkçe)

- [Sentetik tüketici araştırması: sentetik anket ve veri](https://synthetic-consumers.com): Ürün, fiyat ve reklam fikirlerinizi sentetik tüketicilerle test edin. Sentetik anketler ve tüketici dijital ikizleriyle dakikalar içinde araştırma yapın.
- [Demo: bin sentetik tüketiciye bir soru sorun](https://synthetic-consumers.com/demo): Bir soru sorun, bin sentetik tüketici cevaplasın. SCL panelinin küçültülmüş hâlini deneyin.
- [Erken Erişim](https://synthetic-consumers.com/erken-erisim): SCL erken erişim: araştırma sorunuzu paylaşın, sentetik tüketicilerle ilk çalışmanızı birlikte tasarlayalım ve demoyu ilk deneyenlerden olun.
- [Fiyatlandırma: sentetik tüketici araştırması](https://synthetic-consumers.com/fiyatlandirma): Sentetik tüketici araştırmanızın maliyetini hesaplayın. Kişi-Soru başına fiyatlandırma, ücretsiz başlangıç hesabı ve kurumsal paketler.
- [Hakkımızda: sentetik tüketici araştırma ekibi](https://synthetic-consumers.com/hakkimizda): Synthetic Consumer Lab, saha çalışması yerine tüketici dijital ikizleriyle araştırma yapan İstanbul merkezli bir ekiptir. Ekibimizi ve yaklaşımımızı tanıyın.
- [Nasıl çalışıyor? Sentetik tüketici araştırması](https://synthetic-consumers.com/nasil-calisiyor): Sorudan hedef kitleye, simülasyondan rapora: SCL ile sentetik anket ve tüketici araştırması hazırlamanın dört adımı ve sonuçların nasıl okunacağı.

## Yayımlanmış çalışmalar (Türkçe)

- [Çalışmalar: sentetik tüketicilerle araştırmalar](https://synthetic-consumers.com/calismalar): Sentetik tüketicilerle yürütülen açık araştırmalar: atıştırmalıkta fiyat duyarlılığı ve kulaklıkta sosyal kanıtın ödeme istekliliğine etkisi.
- [Atıştırmalık ürünlerinde fiyat duyarlılığı](https://synthetic-consumers.com/calismalar/atistirmalik-fiyat-duyarliligi): İki atıştırmalık üründe fiyat kabulü, talepteki sert düşüşler ve segmentlere göre ödeme istekliliği: sentetik tüketicilerle yapılmış fiyat duyarlılığı çalışması.
- [Çikolata Kaplı Fındıklı Gofret — Fiyat duyarlılığı](https://synthetic-consumers.com/calismalar/atistirmalik-fiyat-duyarliligi/cikolata-kapli-findikli-gofret): Çikolata Kaplı Fındıklı Gofret: test edilen fiyat aralığında beklenen talep, en iyi test fiyatı ve segmentlere göre ödeme istekliliği. Sentetik tüketicilerle yapılmış çalışma.
- [Çoklu Paket Kakaolu Sandviç Kek — Fiyat duyarlılığı](https://synthetic-consumers.com/calismalar/atistirmalik-fiyat-duyarliligi/kakaolu-sandvic-kek): Çoklu Paket Kakaolu Sandviç Kek: test edilen fiyat aralığında beklenen talep, en iyi test fiyatı ve segmentlere göre ödeme istekliliği. Sentetik tüketicilerle yapılmış çalışma.
- [Bir ürünü satın alırken gerçekten neye güveniyoruz?](https://synthetic-consumers.com/calismalar/kulaklik-wtp-karsilastirma): Aynı kulaklık, dört farklı sosyal kanıt senaryosu: puan ve yorumların ödeme istekliliğini (WTP) nasıl değiştirdiğini sentetik tüketicilerle ölçtük.
- [Değerlendirme bilgisi yok — WTP Araştırması](https://synthetic-consumers.com/calismalar/kulaklik-wtp-karsilastirma/degerlendirme-yok): Değerlendirme bilgisi yok — WTP Araştırması: bu senaryoda kulaklık için fiyat kabul oranları, talep eğrisi ve ödeme istekliliği. Sentetik tüketicilerle yapılmış çalışma.
- [Olumsuz yorum — WTP Araştırması](https://synthetic-consumers.com/calismalar/kulaklik-wtp-karsilastirma/olumsuz-yorum): Olumsuz yorum — WTP Araştırması: bu senaryoda kulaklık için fiyat kabul oranları, talep eğrisi ve ödeme istekliliği. Sentetik tüketicilerle yapılmış çalışma.
- [Yüksek puan + az değerlendirme — WTP Araştırması](https://synthetic-consumers.com/calismalar/kulaklik-wtp-karsilastirma/yuksek-puan-az-degerlendirme): Yüksek puan + az değerlendirme — WTP Araştırması: bu senaryoda kulaklık için fiyat kabul oranları, talep eğrisi ve ödeme istekliliği. Sentetik tüketicilerle yapılmış çalışma.
- [Yüksek puan + çok değerlendirme — WTP Araştırması](https://synthetic-consumers.com/calismalar/kulaklik-wtp-karsilastirma/yuksek-puan-cok-degerlendirme): Yüksek puan + çok değerlendirme — WTP Araştırması: bu senaryoda kulaklık için fiyat kabul oranları, talep eğrisi ve ödeme istekliliği. Sentetik tüketicilerle yapılmış çalışma.

## Araştırma kütüphanesi (Türkçe)

- [Makaleler: LLM’lerle tüketici simülasyonu](https://synthetic-consumers.com/makaleler): Büyük dil modelleriyle sentetik katılımcı ve tüketici simülasyonu üzerine akademik çalışmaların özetleri ve öne çıkan bulguları.
- [Analyzing Brand Perception in LLMs](https://synthetic-consumers.com/makaleler/analyzing-brand-perception-in-llms): LLM'lerin markalara yönelik yanlılık (bias) ve tutarlılık düzeylerinin değerlendirilmesi.
- [Beyond Demographics: Aligning Role-playing LLM-based Agents Using Human Belief Networks](https://synthetic-consumers.com/makaleler/beyond-demographics-aligning-role-playing-llm-based-agents-using-human-belief-networks): Demografik bilgiler ve inanç ağları kullanımıyla LLM ajanlarının insan görüşlerini taklit doğruluğu.
- [Can LLMs Capture Human Preferences?](https://synthetic-consumers.com/makaleler/can-llms-capture-human-preferences): LLM'lerin insan tercihlerini taklit etme yeteneği; dil ve chain-of-thought yöntemlerinin etkisi.
- [Comparing Traditional and LLM-based Search for Consumer Choice: A Randomized Experiment](https://synthetic-consumers.com/makaleler/comparing-traditional-and-llm-based-search-for-consumer-choice-a-randomized-experiment): LLM tabanlı aramanın verimlilik, doğruluk ve kullanıcı algısına etkisi; güven vurgularının rolü.
- [Do LLM Agents Exhibit Social Behavior?](https://synthetic-consumers.com/makaleler/do-llm-agents-exhibit-social-behavior): LLM'lerin sosyal etkileşim davranışları; grup kimliği, karşılıklılık mekanizmaları ve CoT süreçlerinin rolü.
- [Enhancing LLM Decision-Making with Factor Profiles and Analogical Reasoning (DEFINE)](https://synthetic-consumers.com/makaleler/enhancing-llm-decision-making-with-factor-profiles-and-analogical-reasoning-define): Belirsizlik altında rasyonel karar; faktör profilleri ve tarihsel analojilerle doğruluğun artırılması.
- [Large Language Models as Simulated Economic Agents: What Can We Learn from Homo Silicus?](https://synthetic-consumers.com/makaleler/large-language-models-as-simulated-economic-agents-what-can-we-learn-from-homo-silicus): LLM'lerin davranışsal ekonomi deneylerindeki insan ajanlarını taklit etme kapasitesi; endowment etkisi ve yeniden üretilebilirlik.
- [Large Language Models as Subpopulation Representative Models: A Review](https://synthetic-consumers.com/makaleler/large-language-models-as-subpopulation-representative-models-a-review): Alt-popülasyon temsil modelleri olarak LLM kullanımı, toplumsal görüşleri ölçme potansiyeli ve risk-fayda değerlendirmesi.
- [Large Language Models Empowered Agent-Based Modeling and Simulation: A Survey and Perspectives](https://synthetic-consumers.com/makaleler/large-language-models-empowered-agent-based-modeling-and-simulation-a-survey-and-perspectives): Ajan tabanlı modelleme ve simülasyonda LLM rolü; simülasyon paradigmasını dönüştürme kapasitesi.
- [LLM-Based Doppelgänger Models: Leveraging Synthetic Data for Human-Like Responses](https://synthetic-consumers.com/makaleler/llm-based-doppelganger-models-leveraging-synthetic-data-for-human-like-responses-in-survey-simulations): Sentetik veriyle bireysel görüşlerin öğrenilmesi ve taklit edilmesi; doğruluk ve heterojenlik.
- [LLMs and Generative Agent-Based Models for Complex Systems](https://synthetic-consumers.com/makaleler/llms-and-generative-agent-based-models-for-complex-systems): GABM'lerle karmaşık sosyal dinamiklerin modellenmesi; insan benzeri davranış taklidi.
- [LLMs’ Ways of Seeing User Personas](https://synthetic-consumers.com/makaleler/llms-ways-of-seeing-user-personas): Kullanıcı personalarının algılanması; kültürel bağlamın etkisi ve demografik profilleri yeniden oluşturma kapasitesi.
- [MatrAIx: Simulating the World with 8.3 Billion Persona Agents](https://synthetic-consumers.com/makaleler/matraix-simulating-the-world-with-8-3-billion-persona-agents): 8,3 milyar persona kaydı ve 1.010 hazır görevle yapay zekâ sistemlerini ve dijital ürünleri popülasyon ölçeğinde test eden simüle kullanıcı değerlendirme altyapısı.
- [Reduce Disparity Between LLMs and Humans: Optimal LLM Sample Calibration](https://synthetic-consumers.com/makaleler/reduce-disparity-between-llms-and-humans-optimal-llm-sample-calibration): Human Mimicry Calibration (HMC) yöntemiyle demografik gruplara göre kalibrasyon.
- [Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization](https://synthetic-consumers.com/makaleler/two-tales-of-persona-in-llms-a-survey-of-role-playing-and-personalization): LLM'lerde persona kullanımı; role-playing ve personalization ayrımı, uygulama alanları.
- [Using Large Language Models to Simulate Multiple Humans and Replicate Human Subject Studies](https://synthetic-consumers.com/makaleler/using-large-language-models-to-simulate-multiple-humans-and-replicate-human-subject-studies): LLM'lerin klasik deneyler ve farklı demografik profiller üzerinden insan davranışlarını simüle etme kapasitesi.