Panel Studio: Run a 1,000 Person Focus Group in Five Minutes

The problem with traditional research

A traditional focus group takes 4-6 weeks to recruit, costs £8,000-15,000, and produces 8-12 opinions. A national survey panel of 1,000 respondents costs £20,000-50,000 and takes 2-4 weeks for fieldwork alone. By the time results arrive, the product decision they were supposed to inform has already been made on gut instinct.

The problem is not that research is bad. The problem is that research is slow and expensive enough that most decisions never get researched at all.

What Panel Studio does

Panel Studio builds synthetic consumer panels from personality science. Each persona is a census-weighted individual with a full demographic profile, an eight-dimension behavioural model (DYNAMICS-8), political beliefs, media consumption habits, and a life history. They are not templates or archetypes. They are computationally constructed individuals whose responses are grounded in the same personality dimensions that drive real human behaviour.

You define a stimulus (a product concept, a policy proposal, an advertisement, a pricing change) and the panel responds. Each persona produces a reasoned, first-person response that explains why they react the way they do. The reasoning traces are auditable: you can read exactly which personality dimensions, life experiences, and demographic factors drove each response.

The economics

This is not a marginal improvement. It is a structural change in the cost of knowing what your customers think.

How the personas work

Every persona is built in three passes:

1. Demographic skeleton: Census-weighted sampling ensures the panel matches real population distributions for age, gender, ethnicity, region, education, income, and household composition. 20 country-specific builders handle national demographic structures.
2. Personality and biography: DYNAMICS-8 scores are assigned based on demographic correlations established in personality science literature. A full biography is generated: childhood, education, career, relationships, beliefs, media habits. This is the context window that grounds every response.
3. Validation: 18 internal consistency rules check that each persona is demographically coherent, behaviourally consistent, and diverse across the panel. Panels are rejected and regenerated if diversity thresholds are not met.

DYNAMICS-8: the behavioural model

DYNAMICS-8 is an eight-dimension personality framework designed for behavioural simulation. Each dimension is a continuous float from 0.0 to 1.0 with four granular facets, giving 32 behavioural parameters per persona.

The framework extends established personality science (Big Five, HEXACO) with two dimensions built for digital behaviour: Acuity and Impulsivity. Published as an open specification under Creative Commons Attribution 4.0.

Reasoning traces: why, not just what

Every persona response includes a full reasoning trace that shows which personality dimensions, demographic factors, and life experiences influenced the response. This is not a black box. You can read exactly why persona #347 (a 58-year-old retired teacher in Swansea with high Discipline and low Impulsivity) rejected your pricing change while persona #891 (a 24-year-old software developer in Manchester with high Novelty and high Acuity) accepted it.

These traces are exportable as structured data for downstream analysis, fine-tuning, or audit.

Multi-turn conversations

Panel Studio supports multi-turn conversations. Present a stimulus, read the responses, then follow up. Ask the panel why they said what they said. Present a revised concept. Test a different price point. Each turn builds on the previous context, just like a real focus group but with 1,000 participants instead of 8.

What you can do with it

• Product testing: test concepts, features, pricing, packaging before building
• Brand research: sentiment tracking, competitor comparison, positioning
• Policy testing: public reaction to regulatory changes, communications
• UX research: test messaging, onboarding flows, feature prioritisation
• Academic research: replicate survey studies, test hypotheses, generate training data
• Election forecasting: constituency-level voting intention from demographically matched panels
• Conjoint analysis: attribute trade-off studies without real respondent recruitment
• Price elasticity: willingness-to-pay curves from personality-grounded responses

Self-hosted or cloud

Panel Studio is source-available under the Business Source License 1.1. Download from GitHub, run docker compose up, and you have a working system with 500 seed personas. No cloud dependencies. Your data never leaves your infrastructure.

Cloud access is available via the Panel Studio API with tiered pricing (free tier included). Python SDK: pip install kronaxis.

How this compares to alternatives

Getting started

# Self-hosted
git clone https://github.com/kronaxis/kronaxis-panel-studio.git
cd kronaxis-panel-studio
docker compose up -d

# Cloud API
pip install kronaxis

Documentation: kronaxis.co.uk/api. Pricing: kronaxis.co.uk/pricing (free tier available).