lecture 5: digital personalization and recommendations

Question 1

defintion personalisation

Answer 1

= combined use of technology and consumer info to tailor electronic commerce interactions
- increased effectiveness
- inkras engagements en relationship with consumer
- offers personalized experience (personal touch)
- assits in search for a product
- increases much more

Question 2

5 S’s digital marketing objectives

Answer 2

1. sell: using internat as sales tool
2. serve: using internat as customer services tool
3. speak: using internat as communication tool
4. sizzle: using internat as brand-building tool
5. save: using internat as cost-reduction tool

Question 3

implicit vs. explicit data

Answer 3

implicit= based on behavior
explicit= collecte trough from (needs to ask)

Question 4

3 types of implicit data

Answer 4

1. context: return visit amount, type of device
2. behavior: content views, liked, bonded products
3. history: past purchase/email interaction

Question 5

personalisation approaches

Answer 5

1. preferente-based authorization
   - select and set-up preferences
2. group customization
   - recommendation based on preferences of people ‘like’ them
3. individualisation
   - uses patterns of own behavior

Question 6

3 levels of personalisation

Answer 6

1. machine driven (1 to 1): fits well
2. rules-based (segmentation): different content to differing groups
3. ab testing (optimisation): which page works best

Question 7

recommendation systems

Answer 7

provides product advise base on
- user-specific preferences
- users’ shopping history
- choices made by others with similar profile

Question 8

3 stages of recommendation systems

Answer 8

1. understand consumer (collection their info & based on this build a profile)
2. deliver recommendation (match profile to product accurate & presentation)
3. impact of system

Question 9

content-based recommendation system

Answer 9

= based on consumer desired product
+ little info needed
- limited in scope
- shallow analysis
- overspecification
- allows change only after new explicit data
- new-users, no profile

Question 10

collaborative-based recommendation system

Answer 10

= use the opinion of like-minded people to generate recommendation (similar profile, you must like movie as well)
+ more accurate
+ based on larger pool of ratings/purchases
- cold start: much info needed
- lagged new item recommendaiton
- poor for unusual users
- large computing need

Question 11

2 other types of recommendation systems

Answer 11

1. hybrid
2. xxx