# What is SCL?

Explains what SCL is, which questions you can answer with synthetic consumers, the core parts of the app and how to read the results.

SCL (Synthetic Consumer Lab) is a research app that puts your market research questions to synthetic consumers. You define a target audience, choose a research method and get the results as a report.

## What is a synthetic consumer?

A synthetic consumer (persona) is a virtual consumer sampled from Türkiye's population and income statistics (TurkStat) and played by an AI language model (the technology behind chat assistants). Each persona has a name, age, gender, province of residence, education level, income group, socio-economic status (SES) class, household details and personality traits. When answering a question, the model behaves according to this profile. In most methods, the personas in a population are split across several different language models so that no single model's habits dominate the result.

Personas are not records of real people and do not correspond to any real person. The [How synthetic consumers work](/en/docs/getting-started/how-synthetic-consumers-work) page explains how they are created.

## Which questions does SCL answer?

Each type of question has its own research method. The abbreviations in the method names mean the following:

- Willingness to pay (WTP): the highest price at which a persona would still buy the product.
- Conjoint: a preference analysis in which personas see your product and its competitors, with their prices, and choose one.
- MaxDiff (best–worst scaling): personas see features in groups of four and rank each group from most to least important, which measures how important the features are relative to one another.

The table below matches common business questions to the relevant method.

| Topic | Example question | Method |
| --- | --- | --- |
| Price | What price range are consumers willing to pay for my product, and which price maximises revenue? | [Price Analysis (WTP)](/en/docs/research/price-analysis) |
| Price | If I change my price, which competitor will customers switch to? | [Competitive Analysis (Conjoint)](/en/docs/research/competitive-analysis) |
| Price | How much does a feature change the price people are willing to pay? | [Feature Price Impact](/en/docs/research/feature-price-impact) |
| Product | Which features matter more in the purchase decision? | [Priority Ranking (MaxDiff)](/en/docs/research/priority-ranking) |
| Product | Which of two products is preferred, and why? | [Comparison](/en/docs/research/comparison) |
| Messaging and perception | How is a message, concept or brand perceived? | [Multiple-Choice Survey](/en/docs/research/multiple-choice-survey) |
| Advertising | How much of my ad do people watch, how much do they like it, do they link it to my brand, and which scene sticks in their minds? | [Ad Testing](/en/docs/research/ad-testing) |
| Qualitative insight | How do consumers describe a topic in their own words? | [Real-Time Focus Group](/en/docs/research/focus-group) |

To pick the method that suits you, see [Choosing the right method](/en/docs/research/choosing-a-method).

## The three verticals on the home page

When you log in to SCL, the **Choose a vertical** screen opens. This screen is both the home page and the starting point of every new study. The **New Research** item in the sidebar and the SCL logo also bring you here.

A "vertical" here is not the industry your company works in but one of the app's three workspaces. If you are testing a product other than clothing, such as a food, drink, personal care or cleaning product, choose **General Research**; price tests are here too.

| Vertical | Card text | When to choose it |
| --- | --- | --- |
| **General Research** | "General-purpose research with market segmentation." | When you want to use one of the seven methods for any product or category |
| **Fashion** | "Apparel comparison on an archetype-based panel." | When you want to test clothing in a population where each persona is assigned a fashion shopping type (archetype) |
| **Ad Testing** | "Test a video ad with a synthetic consumer panel." | When you want to test an ad; this vertical has only one method, so method selection is skipped |

Populations are tied to a vertical: you cannot use a population created in one vertical in another. For details, see [Verticals](/en/docs/research/verticals).

## Core building blocks

### Population

A population is the group of personas a study runs on. You can create a population to match your target audience yourself, or ask Argus to create it. The population is saved and reused in later studies in the same vertical without being generated again.

When you create it, you choose the number of personas, the research topic and filters such as age, gender, province and SES. The research topic is a short phrase describing your product or category (for example, "Premium coffee subscription"); SCL splits the personas into market segments based on this topic. If you like, you can also give some of the personas a feature you define yourself (for example, "Brews filter coffee at home every day"); you do this on the **Custom Feature** tab of the form. See [Creating a population](/en/docs/populations/creating-a-population) and [Population library](/en/docs/populations/population-library).

### Research method and study

The method determines what is measured, how, and what you need to prepare: a price ladder (the list of prices to test) for Price Analysis, a feature list for Priority Ranking, questions for a Multiple-Choice Survey, two products for Comparison, and your ad for an Ad Test. Setting up and running a method on a population is called a study. For the steps, see [Starting a new study](/en/docs/research/new-study).

### Report

In most methods, a progress screen opens when you start the study. How long a study takes depends on the number of personas and the method; the app does not show a time estimate. You don't have to wait on the screen: the study keeps running even if you leave the progress screen. If you stay on the screen, the report opens automatically when the study finishes.

Every completed study appears as a card on the **Reports** page, which you open from the sidebar. Clicking the card opens the report page, with summary figures, charts and tables. You can split the results by dimensions such as age, SES or market segment; these splits are called breakdowns. **Download PDF** downloads a formatted PDF report, and **Download Excel** downloads an Excel file with all the data. See [Reports](/en/docs/results/reports) and [Reading a report](/en/docs/results/reading-a-report).

### Argus

Argus is SCL's AI research assistant. You open it from the **Argus** item in the sidebar or from the round button in the bottom-right corner of other pages. By chatting with Argus, you can have it set up populations and studies, ask questions about results, have it research prices, news and official statistics on the Turkish market, and get the findings as a report or a presentation.

Before it creates a population or starts a study, Argus shows an approval card; the action runs when you press **Approve** or **Approve and launch**. If you turn on the **Autopilot** toggle above the message box, Argus stops waiting for approval in that conversation: it creates populations and also starts credit-spending studies on its own. Autopilot turned on with the toggle has no credit limit. See [What is Argus?](/en/docs/argus/what-is-argus) and [Approvals and Autopilot](/en/docs/argus/approvals-and-autopilot).

### Credits

Studies in SCL run on prepaid credits. Your remaining credits appear on the **Remaining credits:** line at the bottom of the sidebar. Each method has its own credit formula; credits are calculated from the number of personas and the size of the study. For example, a Price Analysis spends 50 personas × 5 price points = 250 credits.

In most methods, you see the estimated credits in the **Credit Usage** field on the **Research Preview** page before you start the study. Creating a population does not spend credits; credits are deducted from your balance after the study's responses have been generated and saved, and in a focus group they are deducted separately for each response. Studies that Argus starts also spend credits; the **Credits to spend** line on the approval card appears only for Fashion and Ad Testing studies and shows the estimated amount. If you belong to an organisation, you use the organisation's shared credit pool. See [Credits](/en/docs/account/credits).

## How SCL differs from a traditional online panel survey

SCL asks the same kinds of questions as an online panel survey, but the answers come from personas rather than real people.

| Aspect | Traditional online panel survey | SCL |
| --- | --- | --- |
| Respondents | Real people registered with the panel | Personas; none of them corresponds to a real person |
| Fieldwork | Participants are invited and you wait for their responses | No participants are recruited; language models generate the responses, and you follow the study on the progress screen |
| Cost | Participant and fieldwork costs | Credits calculated from the number of personas and the size of the study; shown before you start in most methods |
| Repetition | Every new measurement is a new round of fieldwork | You reuse the same population in new studies without generating it again |
| Nature of the results | Statements from real people | Responses generated by language models |

> [!TIP]
> You don't need two studies to compare two prices: put both on the price ladder of the same Price Analysis. You can also compare two products or two packaging designs in a single Comparison study. Separate studies whose results you will put side by side, however, should run on the same population: two populations created separately consist of different personas, even if they are built with the same filters.

## Reading the results

SCL's results are the output of language models. They show which option is ahead, roughly how large the gap is and which groups diverge; they do not give a precise market forecast such as "42% of the market will buy".

- Responses are not identical when you run the same study again. Before you make a decision based on a difference, run the study once more on the same population; this second study also spends credits. If the difference does not come out in the same direction and at a similar size both times, treat it as noise.
- Not every response comes with a reason. In a Multiple-Choice Survey, personas only tick options. You see personas' explanations in their own words in Comparison, Ad Test and focus group studies.

> [!IMPORTANT]
> For high-stakes decisions, we recommend validating SCL results with real consumers. For details of the limitations, see [Accuracy and limitations](/en/docs/getting-started/accuracy-and-limitations).

## Next steps

- To set up your first study step by step, continue with [Quickstart](/en/docs/getting-started/quickstart).
- For definitions of terms such as population, persona, study, segment and credit, see [Core concepts](/en/docs/getting-started/core-concepts).

## Related pages

- [Interface tour](/en/docs/getting-started/interface-tour)
- [How synthetic consumers work](/en/docs/getting-started/how-synthetic-consumers-work)
- [Accuracy and limitations](/en/docs/getting-started/accuracy-and-limitations)
- [Verticals: General Research, Fashion, Ad Testing](/en/docs/research/verticals)
- [Choosing the right method](/en/docs/research/choosing-a-method)