Strategy question

How do you do customer segmentation?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Whether you have a repeatable framework for segmenting customers rather than a vague sense of demographics.

Clarify the goal of segmentation here, for example targeting marketing spend, prioritizing features, or tailoring pricing for Rakuten's ecommerce base.
Start with the data available: behavioral data like purchase frequency and category mix, then demographic and firmographic data as a supplement.
Choose a segmentation method fit for the goal, RFM (recency, frequency, monetary) for retention work or needs-based clustering for product decisions.
Validate segments by checking they are distinct, sizeable, and actionable, not just statistically separable.
Map each segment to a specific action, such as a loyalty push for high frequency low basket-size shoppers.
Confirm with the interviewer whether they want a B2C or marketplace-seller lens, since Rakuten serves both.

Gives a concrete RFM example: high-frequency low-value shoppers get bundling offers, low-frequency high-value shoppers get personalized recommendations.
Explains actionability as the real test of a good segment, not cluster elegance.
Notes that segments should be revisited periodically since behavior shifts, especially around point-based loyalty programs Rakuten is known for.
Distinguishes segmentation for marketing spend versus segmentation for product roadmap, since the axes differ.

Listing only demographic buckets like age and gender without tying them to behavior or business goal.
Not stating what decision the segmentation is meant to drive.
Treating segmentation as a one-time exercise instead of an ongoing process.

How would you validate that a segment is actionable, not just statistically distinct?
How would this segmentation change for Rakuten's loyalty point ecosystem specifically?