What are you trying to learn?
Start from the decision, not the survey. Review the current workflow availability before starting a study. Pick the group that matches your question and we'll show you when to use each method and the estimation behind it.
We block underpowered designs before fielding
Planned designs run through guardrails in the study wizard before a single respondent sees them: sample-size floors for Van Westendorp and Gabor-Granger, and exposure-based rules for MaxDiff and conjoint. Weak designs get a warning with a remedy; statistically invalid ones are blocked — not fielded with a disclaimer.
What should we charge?
Three complementary ways in: map acceptable price ranges fast, trace a demand curve, or model willingness-to-pay from real trade-offs.
Van Westendorp PSM
When to use: Early-stage pricing when you need an acceptable price range before committing to price points.
- Four cumulative price curves
- PMC / OPP / IDP / PME intersections
- Sample-adequacy flags on every read
Gabor-Granger
When to use: Testing specific candidate price points to estimate demand and the revenue-maximizing price.
- Demand curve estimation
- Revenue projection by price rung
- Optimal price point
Conjoint willingness-to-pay
When to use: Pricing a feature or tier when price only makes sense relative to what else is in the package.
- MNL + Hierarchical Bayes estimation
- Holdout validation
- Willingness-to-pay per attribute level
What matters most to customers?
Rating scales give you a sea of 4s and 5s. These methods force trade-offs, so priorities separate instead of clustering.
MaxDiff (best-worst scaling)
When to use: Ranking 8-20 features, messages, or attributes by forcing best/worst choices instead of ratings.
- MNL + Hierarchical Bayes utilities
- Confidence intervals on scores
- Segment-level comparisons
Kano model
When to use: Sorting features into must-haves, performance drivers, and delighters before sequencing a roadmap.
- Functional / dysfunctional question pairs
- Full evaluation-table classification
- Roadmap sequencing by category
Constant sum
When to use: Measuring relative importance when you want respondents to spend a fixed budget across options.
- Fixed-budget point allocation
- Share-of-budget importance scores
Which product configuration wins?
Choice-based methods show respondents realistic product profiles and estimate the value of every attribute level from their decisions.
Choice-based conjoint (CBC)
When to use: Deciding feature bundles, tiers, and packaging when attributes trade off against each other and price.
- MNL + Hierarchical Bayes estimation
- MCMC convergence diagnostics (Gelman-Rubin, ESS)
- Holdout validation
- Market simulator + willingness-to-pay
Adaptive CBC (ACBC)
When to use: Larger attribute spaces where a fixed design would exhaust respondents — the exercise adapts to each person.
- Three adaptive stages
- Build-your-own → screening → choice tournament
- Must-have / unacceptable level detection
Which line-up reaches the most customers?
Individual products can each test well and still overlap. Portfolio methods optimize the combination, not the pieces.
TURF analysis
When to use: Choosing which N flavors, features, or SKUs to ship so the combination reaches the most customers.
- Total unduplicated reach & frequency
- Incremental reach per addition
- Best-combination search
Market simulator
When to use: Stress-testing what-if scenarios — new entrants, price moves, spec changes — on top of conjoint utilities.
- Share-of-preference and first-choice models
- Scenario comparison from estimated utilities
Are customers satisfied — and staying?
Tracking metrics are easy to compute and easy to over-read. Ours ship with the uncertainty attached.
NPS with confidence intervals
When to use: Tracking loyalty over time — with intervals, so a 3-point move on n=80 doesn't get read as a trend.
- Promoter / passive / detractor breakdown
- Wilson-score confidence intervals
- Segment cuts
CSAT & CES
When to use: Measuring satisfaction with a touchpoint (CSAT) or how much effort a task required (CES).
- Registered question types with aggregate scoring
- Works alongside NPS in tracking studies
How we report the numbers
Confidence intervals, sample-adequacy notes, and validation checks travel with every result — see the statistical standards behind the methods.
Statistical honestyLearn the methods in depth
Step-by-step guides on pricing research, MaxDiff, conjoint, Van Westendorp, NPS, and survey design.
Browse the learning centerMatch the method to your decision
Review a suitable starting workflow and its availability, or talk through your research question with our team.