A Practical Guide To Conjoint Analysis
One book, placed in its field — what this book gets right, where it goes further than the rest, and what the rest of the field adds.
Mike West, Editor · Leading With Data · peopleanalyst.com/guides/profile-a_practical_guide_to_conjoint_analysis
Arrow keys or swipe to move · N for notes · F for full screen
A Practical Guide To Conjoint Analysis
One book, placed in its field — what this book gets right, where it goes further than the rest, and what the rest of the field adds.
Mike West, Editor · Leading With Data · peopleanalyst.com/guides/profile-a_practical_guide_to_conjoint_analysis
The problem
Estimated part-worth utilities erodes when it is left to instinct instead of method.
You were taught the moves piecemeal, never the whole model.
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
The model
- 01Experimental Design Quality
- 02Data Collection Method
- 03Consumer Attribute Preferences
- 04Estimated Part-Worth Utilities
- 05Trade-off Valuation
- 06Predicted Market Share
- 07Attribute Importance
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
Step 1 of 5 · Experimental Design Quality
The number-of-levels effect means attribute importance is partly a design choice, so balance level counts deliberately.
Misconception
Practitioners believe that adding more attributes and levels always yields a richer, more precise picture of preference.
Reality
Each extra attribute dilutes respondent attention and each extra level mechanically inflates that attribute's apparent importance through the number-of-levels effect—so more structure often buys you noise and bias, not resolution.
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
Step 2 of 5 · Data Collection Method
Choice-based tasks generally recover trade-offs better than rating scales because they reproduce the purchase moment.
Misconception
Practitioners assume the collection method is a neutral pipe and that choice-based, ratings-based, or ranking tasks yield interchangeable data.
Reality
The task format changes the cognitive process respondents use: rating profiles invites anchoring and straight-lining, while choosing among profiles mimics purchase and captures trade-offs more honestly—so the method partly manufactures the preferences you observe.
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
Step 3 of 5 · Consumer Attribute Preferences
People reliably misreport their own attribute importance, which is precisely why conjoint infers rather than asks.
Misconception
Practitioners believe that if you want to know what customers value, you should simply ask them to rank or rate the importance of attributes directly.
Reality
Direct importance questions produce socially desirable, everything-matters answers because people lack introspective access to their own trade-off weights; conjoint recovers those weights indirectly by watching what respondents sacrifice.
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
Step 4 of 5 · Estimated Part-Worth Utilities
A utility number is only interpretable relative to its own attribute's other levels—never across attributes.
Misconception
Practitioners compare utility values across different attributes—claiming a +1.2 on color beats a +0.9 on price—as if the numbers share a common scale.
Reality
Part-worths are only meaningful relative to other levels within the same attribute; they are zero-centered per attribute, so cross-attribute magnitude comparisons are meaningless without converting to importance or dollar terms.
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
Step 5 of 5 · Trade-off Valuation
Willingness-to-pay is derived arithmetic from utilities, valid only inside your tested attribute and price ranges.
Misconception
Practitioners treat the derived willingness-to-pay as a single hard number that holds across the entire market.
Reality
Trade-off values are interpolations off average utilities and hold only within the price and attribute ranges you tested; they are directional guidance that varies by segment, not a universal price tag.
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
A framework to use
Trade-Off Analysis Framework
This framework is part of the full guide. See the guide →
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
The cheat sheet
- Experimental Design QualityThe number-of-levels effect means attribute importance is partly a design choice, so balance level counts deliberately.
- Data Collection MethodChoice-based tasks generally recover trade-offs better than rating scales because they reproduce the purchase moment.
- Consumer Attribute PreferencesPeople reliably misreport their own attribute importance, which is precisely why conjoint infers rather than asks.
- Estimated Part-Worth UtilitiesA utility number is only interpretable relative to its own attribute's other levels—never across attributes.
- Trade-off ValuationWillingness-to-pay is derived arithmetic from utilities, valid only inside your tested attribute and price ranges.
- Predicted Market ShareSimulated share is share of preference within your modeled set, not projected sales volume.
- Attribute ImportanceImportance is a percentage of tested utility range, so a wider tested range mechanically raises an attribute's importance.
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data
Sources
- A Practical Guide To Conjoint Analysis
A Practical Guide To Conjoint Analysis · Mike West, Editor · Leading With Data