Printable Handouts

PDF

Navigable Slide Index

1. Introduction
2. Source of volume and cannibalization
3. Product and line optimization questions
4. Quick definitions
5. Red bus / blue bus example
6. What is IID? and IIA?
7. Variables we measure and variables we don't
8. We know IIA/IID gets source-of-volume wrong
9. When should we care about substitution patterns?
10. Is it really a problem?
11. Source of volume calculations
12. Cannibalization = Preference
13. Haven't we solved this one?
14. How do you "handle" IIA in your choice models?
15. Simulating preference shares: state-of-the-art
16. What does proportional substitution look like?
17. We have a problem
18. Industry survey: how do current methods do?
19. What do we need?
20. The SoV Probit (1)
21. A very brief introduction to Probit models
22. Why should I care about correlated errors? (1)
23. Why should I care about correlated errors? (2)
24. Why aren't we all using Probit models?
25. The SoV Probit (2)
26. A structured covariance Probit
27. Distance metric 1
28. Distance metric 2
29. Distance metric 3
30. Distance metric 4
31. Choice experiment
32. Example conjoint task
33. Model fit statistics
34. Parameter estimates
35. WTP comparison
36. Representative correlation matrix
37. Does it work? - For a representative individual
38. Does it work? - For a heterogeneous population
39. Success
40. Summary and implications
41. Thank you!

Topics Covered

• Source of volume and cannibalization
• Product and line optimization questions
• Quick definitions
• Red bus / blue bus example
• What is IID? and IIA?
• We know IIA/IID gets source-of-volume wrong
• When should we care about substitution patterns?
• Source of volume calculations
• Cannibalization = Preference
• How do you "handle" IIA in your choice models?
• Simulating preference shares: state-of-the-art
• What does proportional substitution look like?
• Industry survey: how do current methods do?
• The SoV Probit
• Why should I care about correlated errors?
• Why aren't we all using Probit models?
• A structured covariance Probit
• Distance metrics
• Choice experiment
• Model fit statistics
• Parameter estimates
• WTP comparison
• Representative correlation matrix
• Does it work?

Links

Series:

• Bayesian Analysis in Marketing

Categories:

• Marketing & Sales

Talk Citation

Publication History

• Published on March 16, 2010