Use cases
Causal experiments answer one shape of question: given a set of options, which features change the decision, and by how much. That shape recurs across domains, which is why the same platform serves a pricing question and a vaccine acceptance study.
What teams run
| Area | Questions it answers |
|---|---|
| Pricing | Price sensitivity and elasticity. Willingness to pay per feature. Competitive pricing against a named set of rivals. |
| Product development | Which features drive choice and which are table stakes. Where the gaps are. Which improvements are worth building. |
| Market segmentation | Which segments exist by preference rather than by demographics. What each one actually wants. |
| Brand and positioning | Which claims move choice, and for whom. Whether a message that works for one segment repels another. |
| Marketing effectiveness | Which channels and campaigns change behaviour rather than correlate with it. |
| Market structure | Simulated share across competing profiles, including cannibalisation of your own options. |
| Retention and churn | What drives customers to stay or leave, and what a retention offer is worth. |
| Product launch | Market reaction before launch. Go-to-market choices tested rather than argued. |
| User experience | Which design and flow choices change conversion. |
| Public policy | Response to regulation, campaigns, and interventions: vaccine acceptance, immigration preferences, transport choice. |
| Healthcare | Patient preferences, treatment trade-offs, health policy interventions. |
| Sustainability | Attitudes to sustainable options, and what people will trade to get them. |
| Risk and market entry | Entry evaluations, demand estimates, and where the downside sits. |
| Innovation and R&D | Testing ideas before they are built. |
Where it is used
Market research · retail and e-commerce · healthcare and pharmaceuticals · technology and SaaS · media and entertainment · financial services and fintech · public policy and government · education and EdTech · FMCG and CPG · automotive, including electric vehicles · travel and hospitality · telecommunications · energy and utilities.
The academic conjoint literature is largely policy work, which is why the replication record is heaviest there. The same designs run unchanged for a pricing study.
When not to use it
- When you do not yet know the options. A conjoint design needs attributes to test. Use qualitative research first.
- When the decision is not a choice among alternatives. The method measures trade-offs between profiles.
- When you need an absolute number. The output is a relative effect, not a sales forecast.
Evidence for your domain
The human baselines record covers twelve published studies reproduced on the platform, each with its rank correlation against the original human result. That is the most direct evidence for whether the method holds in an area you care about.