Bring your audiences to life

Every brand makes hundreds of decisions a year that need customer input but can’t justify a $10K study.New copy. A pricing change. Three concepts in development. A campaign about to go live. Real research takes 3 to 6 weeks and costs $5K to $200K. So most of those decisions get made without any customer signal at all.

‍

Fairgen Twins sits in that gap.

‍

It’s a marketplace of digital twins: simulated versions of real, named respondents built from premium panel data. You pick an audience, run a study, and get a full research deck in 20 minutes. Quantitative results, qualitative themes, direct quotes from individual twins, and a recommendation summary.

‍

Or skip the structured study. Pick any twin in the audience and talk to them directly. Ask follow-up questions, push back on their answers, go wherever the conversation goes. It works like a depth interview, without the scheduling.

‍

Teams with their own research data can also build private audiences. Upload your own quant and qual research (interviews, surveys, reports), and Fairgen turns it into a set of simulated respondents you own and can query anytime.

‍

This is directional research, not a replacement for field studies. We say that upfront. Twins is for the decisions before the big study, and for the hundred smaller decisions that will never get one.

A simulated version of a real survey respondent.

‍

We take individual-level data from a quantitative study (each row is a real person) and build a detailed profile for each one: what they buy, what they believe, how they respond in their category, how they write and reason (from interview transcripts, if you provide them). Each twin gets an assigned name and photo.

‍

The panel behaves like the audience it was built from. Not like a generic population. That’s the point.

ChatGPT is trained on the whole internet. Ask it what a Gen Z beauty consumer in urban France thinks about your new moisturizer and it gives a confident, smooth, entirely unverifiable answer, optimized for plausibility rather than truth.

‍

Our twins are the opposite. Each one is built from real individual-level data specific to that person, that category, that geography. They disagree with each other. They have different views. They answer the way the underlying research data says they should.

‍

There are no invented profiles. Nothing exists in the platform until real panel data is used to build it.

Directional, not definitive. Every twin is built from real panel data, the same trusted sources researchers already rely on for quantitative studies. That grounding is what separates Fairgen Twins from an AI making educated guesses. The twins reflect how a real, known audience actually thinks and behaves, because that’s what they were built from.

‍

Category-specific twins built from deep research give much stronger results than a generic profile study. The more focused the original data, the more accurate the twin.

‍

The right framing: use Twins to pressure-test before you go to field, not instead of going. Filter three concepts down to one. Find out which message lands. Validate your pricing hypothesis. Then spend the real budget on what you already know will work.

‍

There are things Twins is not the right tool for: brand tracking, segmentation, complex statistical modeling, conjoint analysis. Those need real respondents. Real respondents are also how we build and validate the twins in the first place.

Product managers, marketers, and insights teams who need customer input on decisions that can’t justify a $10K study. Also research agencies looking to extend the life of work they’ve already delivered.

‍

It’s not the right tool for decisions that need statistical confidence. Those go to field.

Five categories of research you can run on Fairgen Twins:

• Concept testing: rank ideas, identify which one wins and why, before committing to fieldwork.
• Messaging & ad testing: test copy, creative, and CTAs before spending media budget.
• Customer discovery: understand pain points and jobs-to-be-done without running a focus group.
• Pricing & packaging: find out how your target segment thinks about price before you set it.
• Brand perception: see how your brand sits in the category and what needs to shift.

‍

For each, the flow is the same. You run a full end-to-end study. The platform helps you build the questionnaire, runs it across your chosen audience, and produces an executive insights deck: quant metrics, qualitative themes, direct quotes, and a recommendation summary. The whole thing takes about 20 minutes.

‍

Or skip the structured study. Pick any twin in the audience and have a one-on-one conversation. Ask follow-up questions, challenge their answers, go deeper on anything they said. It works like a depth interview, without the scheduling.

No. You can start with the marketplace immediately.

‍

The marketplace hosts premium audiences built using premium data. They’re ready to use the moment you sign up. No upload, no setup.

‍

If you want to build a private audience from your own data, that requires uploading your quantitative dataset through the audience creation wizard.

‍

The two options are independent. You can use marketplace audiences today and build private audiences later.

Marketplace audiences are built with data providers, using custom and syndicated data. They are available immediately. No upload required.

‍

Private audiences are built from your own data. Upload a dataset, optionally add transcripts and reports, and get a simulated audience visible only to your account. Build once, test endlessly.

End to end research takes minutes. From picking your audience to running a study to receiving a full executive deck with quant results, qualitative themes, direct quotes, and a recommendation summary, the whole thing is done in under 20 minutes. From there you can go straight into chat with any individual twin to dig deeper on anything the study surfaced.

Two paths, both built for how agencies and panel companies actually work.

‍

The first is the marketplace. If you’ve run deep research on a category or audience, you can publish it as a twin panel on the Fairgen Twins marketplace. Other researchers pay to run studies on it. Your existing data keeps generating revenue, without any additional fieldwork on your end.

‍

The second is private audiences as deliverables. Upload a client’s dataset, build a twin panel from it, and hand it back as part of your research package. The client gets a report and an interactive audience they can keep testing against throughout the year. It extends the life of every study you deliver.

‍

White-labeling is built in. Your logo, your brand colors, your deck. Clients see your branding, not Fairgen’s.

The twin uses its full profile to answer. The richer the training data, the more grounded the answer.

‍

A twin built from a detailed U&A study covering category attitudes, brand usage, and purchase drivers has real context when you ask about a new concept in that category. It won’t just agree. It responds as that specific person would, based on what we actually know about them.

Check the marketplace first.

‍

New audiences are added every week across B2C and B2B categories and geographies. If you need a segment you haven’t researched yourself, a partner may have already built it. Worth a look before assuming you need to start from scratch.

‍

If the audience genuinely doesn’t exist in the catalog and you have no primary data on that group, Fairgen can’t simulate them. We build twins from real respondents. You’d need to collect that data first. Once you have it, the audience lives on the platform indefinitely.

At minimum one core quantitative dataset. A table where each row is a respondent and each column is a question. Behavioral data, brand usage, category attitudes, demographic profiling. The more you include, the stronger the twins.

‍

Interview transcripts are the most valuable thing to add. They give the twins voice and qualitative depth, how these people actually reason, not just what they selected in a grid.

‍

Segmentation decks and analysis PDFs also help. If you’ve already profiled each segment, upload the deck. We map it to the quantitative data.

‍

No strict sample size minimum. A typical U&A or segmentation study is a solid starting point. The audience creation wizard walks you through the mapping step by step.

You do. It’s used only to build twins for that specific audience. It’s never mixed with anything else or used for any other purpose.

‍

Worth checking before you start: if you’re an agency, the data often legally belongs to your client, not you. You’d need their sign-off. We can help you think through how to frame it.

Study mode does this. You sample as many twins as you need, send structured questions across all of them, and get quantitative and qualitative results back as a full report. That covers what most people use a group test for.

‍

A focus-group format (multiple twins in one live conversation) is on the roadmap. Not in this release.