Customer Intelligence - Summative

A clothing store wants to bring back people who used to spend a lot but haven’t bought in a while. Who should get a special 15% off coupon first?

• People who just bought last week

• People who spent a lot before but haven’t bought in 6 months

• People who only buy during big sales and bought last week

• A random 10% of everyone

A movie app wants to spot who might cancel soon. Which sign is strongest?

• The app added more movies

• The person has waited much longer than usual to watch something

• They opened the app email this week

• They tried a new movie genre once

A new user joins a shopping site. The site knows nothing about them yet. What’s the best way to start good recommendations?

• Show only top sellers to everyone for months

• Ask 3–5 quick preference questions during sign-up

• Wait until they buy 5 items

• Show nothing until you have data

A bank wants to help callers during support calls with the next best step. What’s the best plan?

• Always offer the most expensive product

• Rotate offers alphabetically

• Show offers that fit the caller’s needs and are allowed by the rules

• Let the agent guess

A phone company is testing a new save-the-customer script. How do they test it fairly?

• Give it to people based on last name (A–M vs. N–Z)

• Split customers into two random groups: old script vs. new script

• Let agents choose who gets it

• Give it to everyone and compare to last year

Ads report that the last ad someone clicked before buying gets all the credit. A different study shows early video ads also help. Why the mismatch?

• The study must be wrong

• Giving all credit to the last ad ignores earlier ads that helped

• Video ads never help search ads

• Brand search clicks are never helpful

A music app wants to send discounts only to people who wouldn’t renew unless they got a deal. What should they do?

• Send discounts to everyone

• Estimate who changes their mind because of the discount by testing with and without discounts

• Copy the list of last month’s discounters

• Read app reviews for clues

An online grocery store wants one profile per person even if they use different devices. What's the strongest way to match people?

• Match by logins/emails when possible, and use careful guessing only as backup

• Guess by IP and device only

• Cookies only

• Copy-paste into spreadsheets

A shopper asks not to get personalized ads. What should the store do?

• Still use their data for ads but not for reports

• Hide their email but keep using it the same way

• Respect the choice: don’t use their data for targeting and follow the rules

• Delete their account without asking

A store made 120 tiny customer groups and wrote special ads for each. Results are messy and jump around. What likely went wrong?

• Groups are too tiny to learn what works

• They spent too much money

• The ads are too good

• Using a customer data platform caused it