bt_bb_section_bottom_section_coverage_image

Brand Tracking and Consumer Research in Turkey: What Foreign Brands Need to Know

Brand Tracking and Consumer Research in Turkey: What Foreign Brands Need to Know

Brand Tracking and Consumer Research in Turkey: What Foreign Brands Need to Know

Before you commission anything, check which socio-economic classification your Turkish tracker uses, because the national standard changed. In December 2025 the Turkish Researchers Association (Türkiye Araştırmacılar Derneği, TÜAD) published an updated socio-economic status framework that replaces the long-standing two-dimensional model built mainly on occupation and education with a multi-dimensional, household-based model that adds income and economic status indicators. It was built from 21,000 address scans and 6,300 face-to-face household interviews across 26 regions, and weighted to TÜİK’s new urban–rural classification based on 2024 population data.

That matters because SES bands are how almost every Turkish tracker defines its audience and weights its sample. If your brand health study still reports against the old bands, your target definition, your weighting and your trend line all rest on a superseded framework. This article covers that change, what Turkish media measurement actually measures, the official data you should exhaust before paying for anything, data collection modes, sampling, KVKK obligations and how to commission a tracker that survives a volatile market.

Table of Contents

Why imported research designs underperform in Turkey

A global tracker template usually arrives in Turkey with three assumptions baked in: that the local agency can field the same sample design, that the demographic breaks translate, and that the resulting numbers are comparable with other markets in the same study. All three need checking.

The sample design assumption fails most often on urbanity. Turkey is highly but unevenly urbanised, and a design that treats it as a single national population will over-represent large cities unless it is explicitly stratified. The demographic break assumption fails on income, which Turkish respondents report unreliably and which Turkish research practice therefore tends to infer from asset ownership rather than ask directly. The comparability assumption fails quietly: a 2% shift in a brand metric in a market where consumer prices are moving at double-digit annual rates may be noise, a real change in behaviour, or an artefact of a changed basket.

None of this makes Turkish research unreliable. It makes an unadapted design unreliable. The fix is to treat the Turkish wave as a locally specified study that reports into a global framework, rather than a global study executed locally.

The SES framework changed in December 2025

Turkey’s socio-economic status classification was comprehensively updated and published in December 2025, after roughly four years of work. The update was coordinated by TÜAD and co-funded by a group of industry bodies including Reklamverenler Derneği, Reklamcılar Derneği, TİAK, IAB Türkiye, RİAK, ARVAK and MMA Türkiye. The methodology committee drew members from TİAK, TÜAD, Ipsos, SIA Insight and Diyalog.

What changed, and why it affects your tracker:

  • From two dimensions to many. The previous model rested mainly on occupation and education. The new model is household-based and adds income and economic status indicators, producing an SES score per household rather than a lookup from two variables.
  • Income is inferred, not asked. Income is derived from tangible asset indicators rather than self-reported earnings, and occupation is asked in stages rather than as a single question. If your global questionnaire asks household income directly, it is not producing what the Turkish framework uses.
  • Urbanity is built in. Sampling blocks, universe values and weighting use TÜİK’s urban–rural classification based on 2024 population data, which distinguishes dense urban, mid-dense urban and rural areas.
  • The published groups are A, B, AB, C1, C2, D and DE.

The first results cover dense urban areas, where ABC1 accounts for 43.9% of the population — AB at 21.5% and C1 at 22.4% — with C2 the largest group at 32.5% and DE at 23.7%. Nationally, 67.2% of the population lives in dense urban areas, 15.5% in mid-dense urban and 17.2% in rural areas. Results for the other two were planned for release after TİAK’s database research completed in 2026.

Two consequences. Starting a tracker now, specify the new framework from wave one. Running an existing one, ask your agency in writing which framework each wave used — a mid-study change creates a step in your trend line that looks like a brand event and is not.

Who the research industry is

TÜAD is the professional association for the Turkish research sector and the body that coordinates industry-level standards work, including the SES update above. It publishes Turkish-language versions of ESOMAR guidance, including on the mutual rights and responsibilities of researchers and clients, which gives you a reference point for what a professional supplier relationship should look like.

Alongside it sit the joint industry committees that own particular measurement currencies — TİAK for television audience measurement and RİAK for radio — plus IAB Türkiye on the digital side. These matter commercially because when a media owner or agency quotes audience data to you, the question “which measurement currency is this from” has a specific answer in Turkey, and you are entitled to ask it.

Two practical screening questions for any supplier. Is the agency a TÜAD member, and does it work to ESOMAR standards? And for anything that will be published, what is its position on disclosing methodology: sample size, universe, mode, fieldwork dates and margin of error. An agency that resists publishing those is a problem waiting to happen if your research becomes a PR asset.

What Turkish TV measurement actually covers

If television is in your plan, the measurement universe is narrower than the country, and the specification is public. According to TİAK’s published panel specification, the television audience measurement panel works as follows:

Parameter Specification
Universe Households in the 40 measured provinces that occupy a dwelling and have at least one working television set
Settlement threshold Districts with a population of at least 10,000, per RTÜK regulation
Panel size 4,400 gross and 4,000 net households, around 14,000 residents
Age coverage Individuals aged 5 and over, plus guests who watch during visits — around 750 guest viewers reported per day
Technology Passive PeopleMeter measurement, with remote-control buttons identifying which household member is viewing
Continuity Continuous collection, 365 days a year
Stratification Candidates stratified using estimates from a preliminary database survey, weighted with TÜİK population statistics

The two limits worth internalising are the 40 provinces and the 10,000-population district threshold. Turkey has 81 provinces, and the panel universe excludes smaller settlements altogether. That is a defensible design decision for a commercial measurement currency, but it means Turkish TV audience figures describe a specific universe rather than the national population — and if your category sells disproportionately in smaller towns, television reach data will understate or misdescribe your actual audience. Ask for the universe definition alongside any reach number you are given.

Official data to exhaust before you commission anything

A surprising amount of what foreign brands commission in Turkey already exists for free, published by TÜİK on a regular cycle. Exhausting that first lowers what you need to pay for and gives you an independent benchmark against which to sanity-check agency numbers.

TÜİK’s Household Information Technologies Usage Survey is the obvious starting point for anything digital. Its 2026 results, reported in August 2026, put internet usage at 92.3% among people aged 16 to 74, up from 90.9%, with e-commerce usage at 60%, up from 55.7%. Those are population-level figures from a national statistical office, and they are more defensible in an investment case than a vendor estimate.

Beyond that, TÜİK publishes consumer price indices, labour force statistics, household budget surveys, population and address-based registration data, and the urban–rural classification the new SES model uses. The point is not that official statistics answer brand questions — they cannot tell you what people think of you — but that they establish the denominators, the demographics and the macro context your primary research should be read against. Our guide to sizing and validating the Turkish market before you enter covers using these sources for the market sizing question specifically; this article is about the ongoing measurement that follows it.

Choosing a data collection mode

Mode selection in Turkey is a trade-off between cost, coverage and the kind of question you are asking, and the right answer differs from many European markets.

Mode Strengths Weaknesses in the Turkish context Best used for
Face-to-face, in-home or in-street Reaches lower-SES, older and rural respondents; supports long questionnaires and stimulus Most expensive; slowest; fieldwork access and local permissions need handling Benchmark waves, SES-representative studies, concept testing with stimulus
Telephone (CATI) Broad reach; fast; good geographic spread Falling response rates; limited for visual stimulus; call-screening behaviour Short awareness and usage trackers, quick-turnaround reads
Online panel (CAWI) Cheapest per interview; fast; handles visual stimulus well Panel composition skews urban and higher-SES; panel conditioning Continuous trackers among digitally active audiences, creative testing
Mobile-first survey Highest completion among younger respondents; where most Turkish internet use happens Short questionnaires only; layout constraints on stimulus Short pulse studies, post-campaign reads, in-the-moment questions

Given internet usage above 90% among 16-to-74-year-olds, online fielding is viable for most commercial questions. But viability of the mode is not representativeness of the panel: who is in a specific panel is a separate question from who is online, and it is the one to ask. For brand tracking that needs to speak to the whole population including lower-SES and rural audiences, a face-to-face benchmark wave with online continuous waves between benchmarks is a common and defensible compromise.

Sampling Turkey: regions, urbanity and the frame problem

Three structural features of Turkey should appear explicitly in any sample design, and their absence is the clearest sign of an unadapted global template.

Urbanity, on the current classification. With roughly two-thirds of the population in dense urban areas, 15.5% in mid-dense urban areas and 17.2% in rural areas on the classification underlying the new SES work, a design that does not stratify by urbanity will drift toward the cities where fieldwork is cheapest.

Regional structure. Turkey’s statistical regions differ substantially in income, age profile and category behaviour, and a nationally unstratified sample will be dominated by the Istanbul–Marmara axis. Whether you need regional reportability or only regional representativeness is a deliberate cost decision: reportable regional cells multiply your sample size and your invoice.

Istanbul is not Turkey, and is not one market either. It is large and internally differentiated enough that treating it as a single stratum hides more than it reveals where neighbourhood-level income differences drive behaviour.

On frames, ask directly what sampling frame each mode uses — address-based selection, random digit dialling, panel, or quota by interviewer judgement — and get it in the proposal. Quota sampling by interviewer judgement is common, legitimate and cheap, but it is not probability sampling, and a margin of error quoted on a quota sample is a convention rather than a statistical property. Knowing which you bought changes how hard you lean on a small movement.

Questionnaire language and the translation trap

A questionnaire translated into Turkish and a questionnaire written for Turkish respondents are different instruments, and the difference shows up in your data rather than in your proofreading.

The recurring problems are specific. Agreement scales behave differently across cultures, and Turkish respondents’ use of the extremes of a scale is not automatically comparable with another market’s — so a global norm applied to Turkish scale data can mislead in either direction. Brand and category vocabulary often has an established Turkish usage that differs from the literal translation of your internal terminology, and asking about a category using a term respondents do not use produces low recognition that reads as low awareness. Prompted awareness lists need locally correct brand names, including the Turkish spelling and any local nickname a brand actually carries. And politeness norms mean that direct criticism of a brand is often softened, which argues for indirect and projective questioning where you need honest negatives.

The practical requirement is a cognitive or pilot stage before the first full wave — a small number of interviews reviewed for whether respondents understood each question as intended. It is cheap relative to a wasted wave, and it is the step global templates most often skip.

KVKK: panels, incentives and respondent data

Research is personal data processing, and Law No. 6698 on the Protection of Personal Data (KVKK) applies to it. Four things to settle contractually with your agency rather than assume.

Who is the controller. For most commissioned research the agency processes respondent data and you receive anonymised or aggregated results, which usually makes the agency the controller for respondent identity. That allocation should be written down, not inferred, and it changes if you ask for respondent-level data.

What you actually receive. Aggregated tables and anonymised datasets are a different regulatory proposition from respondent-level records with identifiers. Ask for anonymised data by default, and if you need recontact capability, make the lawful basis and the respondent notice explicit.

Special categories. Health, religious belief, political opinion and similar data are special-category under KVKK with a stricter regime. Categories that routinely touch them — pharmaceutical, health and wellness, some financial products — need the handling agreed before fieldwork.

Transfers and incentives. If respondent data moves to your systems abroad, the transfer needs a lawful basis and documentation. Incentive payments create a separate record of identity and payment details, so ask who holds it and for how long.

One point that catches brands out: recontacting respondents later for marketing is not research, and it falls under the commercial electronic message regime rather than your research consent. Treat any marketing use of a research contact as a new decision with its own compliance path — the general consent framework is in our guide to cookie consent and tracking compliance in Turkey.

The political polling rule and your PR survey

Turkey restricts the publication of political polling around general parliamentary elections, and a brand running a survey-led PR campaign can wander into it without intending to.

Article 61 of Law No. 298 prohibits, in general parliamentary elections and from the start of the election period, the publication or broadcast through print, spoken and visual media of public opinion research, surveys, forecasts and similar material — including telephone-based “mini referendum” formats — in favour of or against a political party or candidate, or in a manner that would influence the voter’s vote, and prohibits their distribution in any form.

Two clarifications matter. The provision is directed at publishing and broadcasting results, not at conducting research. And it concerns political content, so ordinary commercial brand tracking is not its target. The exposure arises when a brand’s published survey strays into political or electorally charged territory during an election period — a “what Turks think about the economy” content piece released during a campaign is a different object from the same piece released outside one. If you publish survey-led content, check the electoral calendar and take a Turkish legal view on anything politically adjacent, and confirm the current wording and scope of the provision, which has been amended over its life.

Designing a tracker that survives a volatile market

A brand tracker exists to detect change, which means it has to hold everything else constant. In a market with substantial price movement and frequent category disruption, that discipline is harder and more valuable.

Five design decisions carry most of the weight. Freeze the instrument. Every question change breaks the series; if you must change one, run both versions in parallel for a wave. Hold the mode and the sample design constant, and if you have to change either, bridge it with an overlap wave — a mode change alone can move a metric more than a year of marketing. Set the wave frequency to your decision cycle, not to the budget: quarterly suits most categories, monthly only if you will act monthly. Report price perception as a tracked metric, because in an inflationary environment it moves independently of your pricing and explains movements in value and consideration that otherwise look inexplicable; the pricing side of that is covered in our guide to pricing strategy in Turkey. Separate what you can control from what you cannot by tracking category-level metrics alongside brand metrics, so a category-wide decline is not read as a brand failure.

Finally, triangulate. A tracker is one signal; retailer sell-out data, search demand, first-party behavioural data and complaint volumes are others, and they are often faster. Our guides to trade and shopper marketing in Turkey and customer loyalty and retention programs in Turkey cover where those sit locally.

What to ask an agency before you sign

Eight questions separate a specification you can defend from one you cannot:

  • Which SES framework will each wave use, and if the study spans the transition, how will you bridge it?
  • What is the universe — age range, geography, urbanity — and how does it differ from the national population?
  • What is the sampling frame and method for each mode, and is it probability or quota based?
  • What is the achieved sample by region, urbanity and SES, and which cells will be reportable?
  • How will weighting be done, against which published population figures, and from what date?
  • Who is the data controller for respondent data, what will I receive, and in what form?
  • What is the fieldwork quality process — back-checks, interviewer validation, straightlining and speeding checks?
  • What will you publish about methodology if I use the findings externally?

Get the answers in the proposal document rather than in an email thread. A tracker is a multi-year commitment whose value is entirely in its comparability, and the specification is the only thing protecting that.

Frequently asked questions

Did Turkey’s SES classification really change, and does it affect my brand tracker?

Yes. TÜAD published an updated socio-economic status framework in December 2025, replacing a largely two-dimensional occupation-and-education model with a multi-dimensional household model that adds income and economic status indicators, built from 21,000 address scans and 6,300 face-to-face interviews across 26 regions. Since SES bands define the audience and the weighting in most Turkish trackers, ask your agency which framework each wave used and how any transition is bridged.

Is online research representative in Turkey?

Internet usage is high — TÜİK’s 2026 Household Information Technologies Usage Survey reported 92.3% among people aged 16 to 74 — so online fielding is viable for most commercial questions. But a specific panel’s composition is a separate question from national connectivity, and Turkish panels typically skew urban and higher-SES. For population-representative brand tracking, a face-to-face benchmark with online waves between benchmarks is a common compromise.

Does Turkish TV audience data cover the whole country?

No. TİAK’s published panel specification defines the universe as television-owning households in the 40 measured provinces, in districts with a population of at least 10,000 under RTÜK regulation, with a panel of 4,400 gross and 4,000 net households and around 14,000 residents aged 5 and over. Turkey has 81 provinces, so ask for the universe definition alongside any reach figure.

Who owns the respondent data in commissioned research?

Settle it in the contract rather than assuming. In most commissioned research the agency processes respondent identity and delivers anonymised or aggregated results to the client, which generally makes the agency the controller for respondent data. If you want respondent-level records, recontact rights or data transferred abroad, each needs its own lawful basis and respondent notice under Law No. 6698.

Conclusion

Research in Turkey is well served by a mature local industry, a professional association that runs genuine standards work, and a national statistical office that publishes more free, usable data than most foreign brands exploit. What goes wrong is almost never supplier quality. It is an unadapted global specification.

Three things to get right. Specify the current SES framework from wave one and ask how any transition is bridged, because that single question protects your trend line. Interrogate the universe behind every number you are given, especially in media measurement, where the published panel specification defines a narrower population than the country. And write the questionnaire for Turkish respondents rather than translating it, with a pilot before you commit to a full wave.

Medyae works with foreign and international brands on the marketing side of entering and growing in Turkey — SEO, digital marketing, social media management, web and production — and the local-market and Turkish-language judgement that makes research findings actionable sits close to that work. For how measurement fits the wider growth picture, our pillar guide to marketing in Turkey for foreign brands is the place to start.

This article is general information about research practice and the regulatory environment for consumer research in Turkey, and the figures cited carry the dates of the sources given. It is not legal advice. Data protection obligations, the scope of restrictions on publishing survey results and research-specific compliance questions are fact-specific and should be confirmed with qualified Turkish counsel and with your research supplier.