MaxDiff

Advanced preference scaling that forces users to make tradeoffs between items, revealing true priorities rather than inflated stated importance when items are presented in isolation.

Process

Key Fields

Question it answers	What features do users actually value most? How do priorities differ by user segment?
Participants & timing	200-500 respondents · 5-10 min per response · 1-2 weeks analysis
AI compatibility	AI runs MaxDiff analysis, generates importance scores, and segments respondents by preference profiles.
Output	Feature importance ranking with confidence intervals, segmentation analysis, roadmap recommendation

Use when

Feature prioritization is a critical roadmap decision
Users claim "everything is important" making stated importance unreliable
Tradeoffs between features need to be understood
Segment preferences are likely to differ substantially

Skip when

Fewer than 6 items to prioritize (simple ranking is sufficient)
Qualitative research already shows clear priorities
Sample size is under 100

Common Mistakes

Too many items

Twenty items in a MaxDiff survey creates too many comparisons and exhausts participants. Limit to 8-16 items that are genuinely in contention.

Not explaining the tradeoff framing

"Which is most important?" feels like "which do you like best?" Make the constraint explicit: "Your budget allows only 3 of these features. Which 3 matter most?"

Ignoring segmentation

Aggregated scores may show feature A slightly ahead of feature B, but segment analysis reveals power users prioritize B while casual users prioritize A. Roadmap decisions depend on which segment you are optimizing for.