Personalization in eCommerce

1
Personalization in e-Commerce

• Dr. Alexandra Cristea
• a.i.cristea_at_warwick.ac.uk
• http//www.dcs.warwick.ac.uk/acristea/

2
1. Contents

• Introduction
• Benefits
• Perspectives
• Ubiquitous Computing

3
Introduction

• E-commerce
• The conducting of business communication and
  transactions over networks and through computers.
• the buying and selling of goods and services, and
  the transfer of funds, through digital
  communications
• Others all inter-company and intra-company
  functions (such as marketing, finance,
  manufacturing, selling, and negotiation)
• B2B business interactions between enterprises
• B2C interactions between enterprise and
  customers

4
Benefits

• First Hello Johnny! syndrome
• Cost as issue
• 2005 onwards Customer-Centric services for CRM
  (customer-relationship-management),
• which can flexibly react to dynamically changing
  market requirements
• Customer Data Integration (CDI) services

5
Amazon
6
Amazon
7
Perspectives Use of adaptation

• Often simple business rules, allowing e.g.,
  administrators to offer discounts on the basis of
  products selected by customers

8
Perspectives Personalized Features

• (e.g., BroadView www.broadvision.com)
• Push system is pro-active
• Pull system relies on the user who requests
  information
• Also
• qualifier matching,
• simple rule-based matching business rules
• E.g., generation of electronic coupons (based on
  previous purchases) that are sent by e-mail to
  each customer who has not purchased goods for a
  while

9
PerspectivesPersonalized Product Recommendations

• Generalized
• Interactive, dynamic taxonomies
• Customer behaviour (customers who bought)
• Item similarity (or correlation)
• Personalized
• Content-based (e.g. content-based filtering past
  and present of user) versus social
  recommendations (collaborative filtering) pros
  cons
• hybrid recommender systems
• Item-to-item collaborative filtering (similarity
  to content based item similarity, but
  lightweight, without user for stable products)

10
Perspectives Customer info sharing

• As a solution to latency (cold start) central UM
• Issues?

11
Perspectives Personalized Product Info

• leading to a sale
• E.g., evaluation-oriented (as a car-sales person)

12
Case Study SeTA

• sorting items on a suitability basis, to the
  preferences of their beneficiary.
• Individual UM (direct questionnaires
  monitoring) indirect (stereotype)
• demographic data (e.g., age, job), preferences
  for products (e.g., products).
• Prologue and summary tailored to user
• User vendor interests represented
• Comparison table is allowed

13
Beginner (non expert)
14
Advanced (expert user)
15
Conclusions Case Study SeTA

• Positive advanced UM, dynamic content generation
  techniques, personalized recommendation
  generation of electronic catalogs meeting
  individual user needs with high accuracy.
• Negative knowledge intensive approach supporting
  the system adaptation which may discourage web
  designer.

16
Perspectives CRM

• customer-centered instead of product-centered
• share of customer, replacing traditional share of
  market.
• accurate UM can then support the proposal of
  personalized offers to improve the customers
  loyalty and thus the companys profit, in the
  medium-long term
• mass customization
• Cross-selling, up-selling

17
Perspectives Mass Customization

• Custom-design (for real!)
• Issues costly (for firm) difficult (for
  customer)
• Adaptation can help with the latter via
  intelligent interaction with the buyer

18
Context-aware and Ubiquitous Computing in
e-Commerce

• accessing a service anytime, anywhere and via
  different types of (mobile) devices.
• M-Commerce commercial transactions performed by
  using wireless devices
• E.g., digital wallets, push information services,
  and location-based services (e.g., visiting a
  museum, or attending a concert, or driving on a
  motorway)
• Issues power, bandwidth, efficiency, screen size
  limitations

19
Ubiquitous m-Commerce Perspectives

• generation of product and service presentations
  whose length is tailored to the screen size.
• layout of the user interface to the
  characteristics of the device used to access the
  service. (via HTML or XML processing, e.g.)

20
Conclusions Discussion

• Here B2C
• Potential personalization also in B2B
• Quality of Service (QoS) levels
• (web) Service discovery, composition, execution
• Web Services description languages, e.g. WSDL
  enable the specification of service public
  interfaces.
• Web Service orchestration languages, e.g.,
  WS-BPEL, support the definition of composite
  services based on the orchestration of multiple
  providers within possibly complex workflows
• Semantic Web techniques have been used to add
  personalization to Web Services

21
Conclusions

• Personalization in e-Business yes, if
• Supporting CRM (cust-rel-mng)
• Enhancing usability
• Enhancing interoperability

22

• Any questions?