baqmar 2008
DESCRIPTION
BAQMaR Conference 2008TRANSCRIPT
![Page 1: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/1.jpg)
![Page 2: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/2.jpg)
![Page 3: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/3.jpg)
![Page 4: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/4.jpg)
600+ analysts
![Page 5: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/5.jpg)
Our dreams
![Page 6: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/6.jpg)
Inspiration
![Page 7: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/7.jpg)
Friend organisations
Marketing research domains
Neighbour countries
Online & offline
![Page 8: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/8.jpg)
website!
![Page 9: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/9.jpg)
Jointhe conversation !
![Page 10: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/10.jpg)
Marketing Researchers TODAY ?!
![Page 11: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/11.jpg)
C L
PopularUnique
Well known
![Page 12: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/12.jpg)
![Page 13: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/13.jpg)
![Page 14: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/14.jpg)
An explosion of data !
![Page 15: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/15.jpg)
![Page 16: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/16.jpg)
Fatigue among respondents !
![Page 17: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/17.jpg)
![Page 18: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/18.jpg)
Me, MySpace & IVisual (n) ethnography among 13-17 year olds
Joeri Van den Bergh (InSites Consulting)
Veerle Colin (MTV Networks)
![Page 19: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/19.jpg)
![Page 20: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/20.jpg)
WHY TELL ME WHY
The Research Briefing
![Page 21: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/21.jpg)
Getting intimate with our target groups
ID construction of adolescents in a
digitalised world
Social groups as an extension of
psychographic segments
Role of brands in ID construction within
social groups
![Page 22: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/22.jpg)
I SEE YOU, BABY
The Research Approach
![Page 23: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/23.jpg)
Hawthorne effect
Researcher
gazeTime & Cost
intensive
Traditional
ethnography
context
Participant
R
![Page 24: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/24.jpg)
Hawthorne effect
Researcher
gazeTime & Cost
intensive
Traditional
ethnography
context
Participant
R
360° visual
ethnography
Contact with researcher
via 2.0 tools“Informer” gaze: what is
important?Follow multiple participants from anywhere
R
Participant
context
From traditional to visual ethnography
![Page 25: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/25.jpg)
1.User generated MM ethnography
= Observation takes place via photos/video
taken by the participants to the study
= Participants observe their own
environment & report back to the
researcher
360° Ethnography
![Page 26: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/26.jpg)
![Page 27: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/27.jpg)
Personal mePictures of...
Place where you can
really be yourself
Clothes you wear at
home
Objects that are
typical for me
Non mePictures of...
Clothes you no
longer want to
wear
Youngsters where
you so not want to
be friend with
Aspirational me
Pictures of...
Favorite clothes
youngsters where
you would like to
be friend with
Social mePictures of...
Important persons
in your life
Clothes you wear
to go out
Groups of
youngsters you like
Other groups
Pictures of other groups that
are different but okay
Pictures of youth of today
Pictures of normal persons
Ethnographic blog: identity related tasks
![Page 28: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/28.jpg)
1.User generated ethnography
= Observation takes place via photos/video taken
by the participants to the study
= Participants observer their own environment &
report back to the researcher
2.Nethnography: social life of teenagers is very
much MOVING ONLINE !!!
= observation of the online behaviour & content
of a target group or within a certain webspace
360° Ethnography
![Page 29: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/29.jpg)
Nethnography: we are your friends
![Page 30: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/30.jpg)
Personal
identity
Nicknam
e
Profile
text
Profile
picture
Photo
collection
Clan member
ship
Conversation
Monitoring on
guestbook
Social
identity
Social Network Sites: nethnography
![Page 31: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/31.jpg)
And for the datamining freaks among you, Annelies ripped the internet
• 300 active participants of netlog randomly
selected. Equal spread age x gender
• Webcrawlers to „scan‟ pages of netlog and
substract content.
• Textmining: profile pages – photo tags –
clan membership – conversations on the
guestbook
![Page 32: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/32.jpg)
Other online behaviour: tracking tool
![Page 33: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/33.jpg)
100% UGC OBSERVED BY US
That‟s why we do call it (visual) (n)ethnograpy: shoot me!
![Page 34: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/34.jpg)
THE HARD PART
The Research Analysis
![Page 35: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/35.jpg)
Analysis
INTIMACY
![Page 36: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/36.jpg)
![Page 37: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/37.jpg)
ME-SEUM
![Page 38: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/38.jpg)
Social Networks are not so social as you think they are
Only 32% of the conversations on the guestbook
are „interactive‟!
The rest are all single statements.
Feedback on picture
Congratulations
School
Confirm friendship
MSN
Express love
Miss you
Practical appointment
It was great
Party
Making fun
Family
Study
How are you?
Age
Music
Online movies
Transport
I’ am bored
Feedback on profiles
Festivals
Sport
Mobile phones
Welcome by strangers
Youth movement Alcohol
Food
Litte kids
Quarrel
Gaming
Food
Sleeping
Tokio hotel
Travel
Cars
Shoes
![Page 39: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/39.jpg)
one of the 3 key research questions
Social groups among today‟s youngsters
![Page 40: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/40.jpg)
To what
youngster
group
do I belong?
![Page 41: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/41.jpg)
Methodology
Aspira
tional
Group
Social
Group
Differe
nt but
OK
Non
group
M
E
![Page 42: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/42.jpg)
WE ME (I’m better)T
HIN
KD
RIN
KCHANGE
CONSERVATISM
REAL WORLD
FANTASY
Fashion girlFashion boy
Breezer sluts
Jumpers
Nerds
Geek girls
emo
Alternative
Punk
Gothic
RockersTektonic
Skater
Hippies
Rapper
MAINSTREA
M
![Page 43: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/43.jpg)
spartacus121 best pk (runescape)
30/04/2008 19:07:55 door laurent
ik hit iets in de 27 en dat is al zot hoog zij hits zijn
gestoord lol. zen zwaard is 140M (ik heb 10 M)
![Page 44: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/44.jpg)
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS SKILLS
SKILLS
SKILLS
SKILLS
SKILLSSKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS SKILLS
SKILLS
SKILLS
SKILLS
SKILLSSKILLS
SKILLS
SKILLS
WE ME (I’m better)T
HIN
KD
RIN
KCHANGE
CONSERVATISM
REAL WORLD
FANTASY
Fashion girlFashion boy
Breezer sluts
Jumpers
Nerds
Geek girls
emo
Alternative
Punk
Gothic
RockersTektonic
Skater
Hippies
Rapper
![Page 45: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/45.jpg)
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS LOOKS
LOOKS
LOOKS
LOOKS
LOOKSLOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS LOOKS
LOOKS
LOOKS
LOOKS
LOOKSLOOKS
LOOKS
LOOKS
WE ME (I’m better)T
HIN
KD
RIN
KCHANGE
CONSERVATISM
REAL WORLD
FANTASY
Fashion girlFashion boy
Breezer sluts
Jumpers
Nerds
Geek girls
emo
Alternative
Punk
Gothic
RockersTektonic
Skater
Hippies
Rapper
![Page 46: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/46.jpg)
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS LOOKS
LOOKS
LOOKS
LOOKS
LOOKSLOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS LOOKS
LOOKS
LOOKS
LOOKS
LOOKSLOOKS
LOOKS
LOOKS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS SKILLS
SKILLS
SKILLS
SKILLS
SKILLSSKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS SKILLS
SKILLS
SKILLS
SKILLS
SKILLSSKILLS
SKILLS
SKILLS
WE ME (I’m better)T
HIN
KD
RIN
KCHANGE
CONSERVATISM
REAL WORLD
FANTASY
Fashion girlFashion boy
Breezer sluts
Jumpers
Nerds
Geek girls
emo
Alternative
Punk
Gothic
RockersTektonic
Skater
Hippies
Rapper
![Page 47: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/47.jpg)
![Page 48: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/48.jpg)
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS LOOKS
LOOKS
LOOKS
LOOKS
LOOKSLOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS
LOOKS LOOKS
LOOKS
LOOKS
LOOKS
LOOKSLOOKS
LOOKS
LOOKS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS SKILLS
SKILLS
SKILLS
SKILLS
SKILLSSKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS
SKILLS SKILLS
SKILLS
SKILLS
SKILLS
SKILLSSKILLS
SKILLS
SKILLS
WE ME (I’m better)T
HIN
KD
RIN
KCHANGE
CONSERVATISM
REAL WORLD
FANTASY
Fashion girlFashion boy
Breezer sluts
Jumpers
Nerds
Geek girls
emo
Alternative
Punk
Gothic
RockersTektonic
Skater
Hippies
Rapper
![Page 49: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/49.jpg)
REMEMBER ME
How this research changed our lives
![Page 50: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/50.jpg)
6 changes that rocked our socks off
• The tools for ID construction have changed
• New online quali methods proved to be efficient
• Our target group = the new and better quali researchers
• New reflection kit for entire MTV Networks staff
• Closer connection with MTV & new clients
• Redefine content strategy of TMF on screen & on line
![Page 51: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/51.jpg)
![Page 52: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/52.jpg)
![Page 53: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/53.jpg)
![Page 54: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/54.jpg)
4C Consulting
Introduction to our services
![Page 55: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/55.jpg)
Our Mission | Boosting your customer value
4C Consulting
helps companies
win, keep and grow
customer value
![Page 56: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/56.jpg)
Our Solutions | Call us for…
Customer Value
Strategy
Process
Excellence
Customer
Insight
Business
requirements
definition
Package selection
& implementation
Post-launch
care
![Page 57: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/57.jpg)
1. Acquire new customers
2. Sell more to existing customers
• More of the same (increase turnover)
• Expand value proposition portfolio
(cross-sell & product development *)
• Upgrade value
proposition (Upsell)
3. Prevent existing
customers from leaving
1. Efficient Delivery (Process
Excellence)
2. Align value propositions
3. Advanced pricing
Our Focus | Boosting your customer value
Increase Revenue Reduce Costs
![Page 58: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/58.jpg)
Business Intelligence Practice
SCV ROMSCCI Competing on Analytics
Audit
Infrastruc-tuur
Data Quality
Exploitatie Coaching
![Page 59: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/59.jpg)
Why 4C Consulting | 7 compelling reasons
1. 100% focus on customer value
management
2. Result-oriented project approach
3. Connecting marketing, sales &
customer care with senior management
and IT
4. Independent consultant for 10 years
5. Experienced crew, passionate about
marketing, sales & customer care
6. Value based pricing model
7. Satisfied & loyal customers: 90
customers, more than 380 projects
![Page 60: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/60.jpg)
60
![Page 61: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/61.jpg)
Optimize your business with Business
&Decision
Michel Meulders
- Domain Manager -
![Page 62: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/62.jpg)
Business & Decision Benelux
Business & Decision Benelux is :
• a multi-specialist, in specific
technology fields :
• Business Intelligence
• Customer Relationship Management
• Life sciences
• Risk & Compliance
• with foreign offices in Brussels,
Amsterdam, Luxembourg
• Top accounts in finance, pharma,
telco, distribution, industry (Fortis, ING,
ABN Amro, Dexia, GSK, UCB, Proximus,
Belgacom, Carrefour, Honda,…)
Founded in 2002
Merger of several companies specialised in BI
Consulting & System Integrator :
- More than 300 consultants
- About 18 mio Euro turnover in 2007
- 58% organic growth comparing to 2006
- Last acquisition : BnV Group (BE+NL)
Turnover evolution (consolidated)
0
5000
10000
15000
20000
25000
2004 2005 2006 2007 Obj 2008
Th
ou
sa
nd
s
Belgium Luxembourg Netherlands
![Page 63: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/63.jpg)
For more info see http://www.businessdecision.com
![Page 64: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/64.jpg)
Belgian Federation of
Market Research Institutes
www.febelmar.be
![Page 65: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/65.jpg)
![Page 66: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/66.jpg)
![Page 67: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/67.jpg)
Febelmar mission
Development and promotion of market
research and opinion polls in Belgium
Protecting the sector interests
Watching over correct use of deontological
rules of market research in all phases of the
market research process
Stimulating continuous improvement of
quality of service in market research
Being a platform for communication,
exchange of expertise and networking
![Page 68: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/68.jpg)
Members
27 agencies.
Together they represent about 75% of
the total market research expenditures in
Belgium.
![Page 69: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/69.jpg)
![Page 70: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/70.jpg)
![Page 71: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/71.jpg)
What is I4BI?
i4bi is specialized in implementing Business Intelligence Solutions in
your company.
Our team of BI experts has deep functional and technical experience
with Application development, Business Model definitions and Data
Warehousing. Functional/technical designs, development, application
role out, training etc… are phases in a project where our consultants
have many years of experience.
i4bi consultants have a deep knowledge of the Oracle Business
Intelligence products and solutions.
i4bi sponsors the development of an independent analytical branch,
which will probably see the light in 2009
![Page 72: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/72.jpg)
What do we provide?
We Provide
Business
Experience
Analytical
Expertise
To Support
Technical
Abilities
Strategic Decision Making
![Page 73: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/73.jpg)
Expertise
• Analytical Expertise• Data Mining
• Statistical modelling
• Predictive analysis
• Basel II compliant modelling
• Forecasting
• Business Analysis Expertise• Reporting
• Delivering Business Insight to decision makers
• Marketing Analysis
• Data Quality
• Use of technical tools such as SAS – SPSS – Statistica to
support & extend business knowledge
![Page 74: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/74.jpg)
Contact
• For more general information:
www.I4BI.be
• For more analytical information:
![Page 75: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/75.jpg)
InSites Consulting
6 beliefs in 60 seconds
![Page 76: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/76.jpg)
“Consumers are beginning in a very real sense to own our brands and participate. We need to begin to learn how to let go”
A.G. Lafley, CEO & Chairman of P&G
We believe ...in the empowered consumer
Human-to-human interactions are more powerful than ever and can make or break your brand
![Page 77: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/77.jpg)
We believe ...in giving back
Rewarding experiences for participantsActive involvement of panel members
Charity contributions
![Page 78: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/78.jpg)
We believe ...in connecting
Everything we do is aimed at strengthening connectionsbetween you, your market and us
“Connected Research” brings you closer to your market and taps into the wisdom of the crowds
Some of our connected research methods
Research communitiesBulletin boardsBlog research
Online discussion groups
More information: http://connectedresearch.insites.eu/
![Page 79: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/79.jpg)
We believe ...in the power of new research methodsfor better marketing decision making
InformationalProviding more depth to research insights
TransformationalDoing things that were previously not possible
AutomationalConducting research more efficiently
![Page 80: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/80.jpg)
We believe ...in 1 + 1 = 3
Old and new methods need to be optimally “fused” in order to fully grasp the new customer / consumer reality
![Page 81: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/81.jpg)
We believe ...in the power of our team
People make the differenceOpen, forward thinking, dedicated, passionate
Specific knowledge centers
![Page 82: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/82.jpg)
©Keyrus – all rights reserved
BAQMaR, 17 December 2008
Welcome to Keyrus
performance management consulting technology
![Page 83: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/83.jpg)
About Keyrus (Belgium)
• founded in 1996 as SOLIDPartners
• focus on performance management, business intelligence & data warehousing
• strong and balanced client base spread over different industries
• +100 consultants specialised in both technical and business domains
• part of Keyrus group (France)
![Page 84: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/84.jpg)
Keyrus‟ global footprint
head office in Paris
present in 9 countries
listed on Paris stock
exchange Euronext
+1300 employees
![Page 85: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/85.jpg)
Vision & mission
Keyrus will be one of the few leading service providers
in the area of performance management.
We help our clients to effectively design, build and operate
the adequate performance management organization and solutions
in an integrated end-to-end fashion.
![Page 86: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/86.jpg)
Portfolio of solutions & services
Information Management
IM
Business Intelligence Platforms
BIP
Analytic Applications
AA
People and Processes
P&P
Corporate Performance Management
(C)PM
data warehouse & data marts
dat
a d
eliv
ery
&
ou
tflo
w f
un
ctio
ns
CPM data delivery, exchange & synchronization
info
rmat
ion
m
anag
em
en
t fu
nct
ion
s
sou
rce
sys
tem
s &
ap
plic
atio
ns
BI L
aye
r (r
ep
ort
s, O
LAP,
d
ash
bo
ard
s, a
lert
s)
CPM Applications (e.g. planning, ABM, PA)
Analytic Applications (e.g. data mining)
![Page 87: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/87.jpg)
Keyrus [email protected]
Nijverheidslaan 3/2 B-1853 Strombeek-Bever
t +32 2 706 03 00f +32 2 706 03 09
Contact us
www.keyrus.be
17-Dec-2008performance management consulting technology
![Page 88: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/88.jpg)
Introduction of Profacts
![Page 89: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/89.jpg)
Who is PROFACTS ?
We are „the new kids on the block‟
in (online) market research ...
![Page 90: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/90.jpg)
Q4
2006
Q4
2008
Q1
2007
Q2
2007
Q3
2007
Q4
2007
Q1
2008
Q2
2008
Q3
2008
2people have
founded Profacts
6people are now
working @ Profacts
200%growth rate
REVEALING FACTORS FOR SUCCESS1 strategy
28yrs
mean age
![Page 91: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/91.jpg)
Profacts is active in more then 10 sectors ...
REVEALING FACTORS FOR SUCCESS
AUTOMOTIVE
GPS
TELECOM ENERGY
ICT PHARMACEUTICAL
FMCG
RECRUITMENT
INSURANCESBANKING
![Page 92: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/92.jpg)
Python Predictions
![Page 93: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/93.jpg)
Python Predictions
PREDICT
![Page 94: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/94.jpg)
Python Predictions
![Page 95: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/95.jpg)
Python Predictions
GROWTH
![Page 96: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/96.jpg)
Python Predictions
www.pythonpredictions.com
![Page 97: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/97.jpg)
Rogil Research
A research agency with a view
![Page 98: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/98.jpg)
MARKETING & SENSORY RESEARCH
OUR PASSION
FTF research (Mobile Unit)
Telephone research
Online research
Panel services
Fieldwork in Europe
Sensory research
Eye-tracking / Eye|watch
Tachistoscope
Trained panel
Consumer panel
Taste lab
![Page 99: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/99.jpg)
![Page 100: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/100.jpg)
Sensory Safari
Note down in your agenda
SENSORY SAFARI
• March 26th 2009
• 18u
• At Rogil in Leuven
![Page 101: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/101.jpg)
Sensory Safari
5 SENSES MARKETING
![Page 102: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/102.jpg)
We hope to welcome you in March.
Thanks for your attention !
![Page 103: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/103.jpg)
SAS Analytics For Challenging Times
Start Focused, Think Wide
![Page 104: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/104.jpg)
Campaign Managment Requires
Optimization
![Page 105: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/105.jpg)
CRM is becoming Risk Managment
![Page 106: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/106.jpg)
SAS Breadth of Analytic Offering…
• Statistical Analysis
• Survey Design/Analysis
• Data Mining
• Text Mining
• Time Series Mining
• Forecasting
• Quality Improvement
• Operations Research
![Page 107: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/107.jpg)
SAS Innovations in Marketing
Solutions….
![Page 108: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/108.jpg)
Copyright © 2006, SAS Institute Inc. All rights reserved.
http://www.sas.com/feature/analytics/index.html
![Page 109: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/109.jpg)
The Mission
Drive the widespread use of
data in decision making
![Page 110: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/110.jpg)
The Focus
Attract Grow Fraud RiskRetain
Driving and Maximizing Profit
![Page 111: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/111.jpg)
The Vision
Behavioral dataDescriptive data
Attitudinal dataInteraction data
Enterprise
Data
Sources
Operational
Processes
Operational
Processes
Enterprise
Data
Sources
![Page 112: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/112.jpg)
The Acceptance
• The rise of the agnostics
Science vs. Chance
In numbers we trust!
![Page 113: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/113.jpg)
![Page 114: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/114.jpg)
![Page 115: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/115.jpg)
![Page 116: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/116.jpg)
The myth of the „best‟ algorithm
lessons learned from innovations in
data sampling and data pre-processing
for marketing analytics
Dr. Sven F. CroneDeputy Director, Ass. Prof.
![Page 117: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/117.jpg)
Associated Experts
Prof. Paul Goodwin
Dr. Andrew Eaves
Research & PhD students
Heiko Kausch, RA
Stavros Asimakopoulos
Xi Chen
Bruce Havel
Suzi Ismail
Nikolaos Kourentzes
Ioannis Stamatopoulos
Andrey Davidenko
Charlotte Brown
Hong Juan Liu
T Hu
John Prest
Huang Tao
Visiting Researchers
Prof. Geoff Allen
Dr. Yukun Bao
Young-Sang Cho
Directors
Prof. Robert Fildes
Prof. Peter Young
Dr.Sven F. Crone
Researchers
Dr. Steve Finlay
Dr. Alastair Robertson
Dr. Didier Soopramanien
Dr. Kostas Nikolopoulos
Prof. Stephen Taylor
Dr. Wlodek Tych
Prof. David Peel
Prof. Peter Pope
![Page 118: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/118.jpg)
“Take away this pudding, it has no
theme.” Sir Winston Churchill (1915)
![Page 119: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/119.jpg)
• Sampling issues in Data Mining
• Case study 1: Direct Marketing• Cross-selling of Magazine subscriptions
• Effect of data preprocessing: Sampling
• Interaction of Sampling with Scaling & Coding
• Case study 2: Credit & Behavioral Scoring• Predicting consumer credit default
• Effects of sample size
• Effects of sample distribution
• Case study 3: Online Shopping Behaviour• Predicting consumer shopping channel choice
• Sample distribution & multiple classes
• Conclusion & Take-aways
Agenda
![Page 120: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/120.jpg)
Why (Under/Over) Sampling?
• Knowledge Discovery (KDD) = non-trivial process of identifying valid, novel, useful patterns in large data sets• Data Mining = only one single step in the KDD process• Data sample determines the whole process! ( GIGO)• “Research seems preoccupied with algorithms” [Hand 2000]
Monitoring
CRISP-DM Process
SAS SEMMA DM-Process
![Page 121: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/121.jpg)
Sampling in Direct Marketing Literature?
Input
type* Methods***
Paramete
r tuning
Data reduction** Data projection
Feature
Selection
Re-
sampling
Continuous attributes Categories
Standardisation Discretisation Coding
[2] 2 BMLP, LR, LDA, QDA X X
[42] 1 MLP, LR, CHAID X X
[43] 2 MLP, RBF, LR, GP, CHAID X X
[44] 3 MLP, LR, LDA X X
[4] 2 CHAID, CART X
[6] 2 MLP, LR X X X X X
[9] 2 LVQ, RBF, 22 DT, 9 SC X X
[45] 2LDA, LR, KNN, KDE, CART, MLP,
RBF, MOE, FAR, LVQX X
[3] 1 MLP X X
[7] 2 LSSVM X X X
[11] 2 LR, LS-SVM, KNN, NB, DT X X X
[10] 1LDA, QDA, LR, BMLP, DT, SVM,
LSSVM, TAN, LP, KNNX X
[46] 2 LR, MLP, BMLP X X
[47] 2LSSVM, SVM, DT, RL, LDA, QDA,
LR, NB, IBLX X
[48] 1 DT, MLP, LR, FC X
[49] 1 FC X X
Majority of direct marketing papers focus on algorithm tuning
Only 3 papers consider Resampling / Instance Selection
No analysis of the interaction with Sampling & Projection & …
![Page 122: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/122.jpg)
Database of customers (instances)
Known attributes for all customers (age, gender, existing subscriptions, …)
Known response (class membership) of buyers & non-buyers from past mailings
Build a model to separate classes decision boundary of different complexity
1 … Number of subscriptions … Many
Classification
Few
… D
ays s
ince last
purc
hase
… M
any
No responseSubscribed to magazine
Last campaign
![Page 123: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/123.jpg)
Class unknown
Use the decision boundary to classify unseen instances
Calculate on which side of hyperplane the instances lie (or distance)
Assign class to unseen instances
No responseSubscribed to magazine
Classification
1 … Number of subscriptions … Many
Few
… D
ays s
ince last
purc
hase
… M
any
![Page 124: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/124.jpg)
Balanced dataset = class distributions are equal P(x|y=A)=P(x|y=B)
proportional sampling or stratified sampling feasible
Imbalanced dataset = class distributions unequal P(x|y=A)>>P(x|y=B) `
The class of interest is often the minority (in most business applications)
Reality Check: Imbalanced classes
No responseSubscribed to magazine
Problem
• Classifiers are biased towards
the majority class
• Shifts the decision boundary
• Error / Accuracy based learning
creates naïve classifiers
• Invalid separation of classes1 … Number of subscriptions … Many
Few
… D
ays s
ince last
purc
hase
… M
any
![Page 125: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/125.jpg)
Size of the sample?
Distribution / location of the sample?
Imbalanced Data Sampling
No responseSubscribed to magazine
Stratified Random Sampling
divide DB in mutually exclusive
strata (subpopulations) & draw
random samples from each
Proportional
assure proportions in samples
equal those in population
Disproportional
weighted over-& undersampling
of important classes1 … Number of subscriptions … Many
Few
… D
ays s
ince last
purc
hase
… M
any
![Page 126: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/126.jpg)
Exclude random instances of the majority class
Retain all instances of the minority class
Establish a balanced class distribution
Random Undersampling
No responseSubscribed to magazine
Benefits
• Helps detect rare target levels
Risks
• Biases predictions (correctable)
• Looses information contained in
instances of the majority class
• Creates different boundaries
• Increases prediction variability
• …1 … Number of subscriptions … Many
Few
… D
ays s
ince last
purc
hase
… M
any
![Page 127: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/127.jpg)
Retain all instances of the majority class in the sample
Duplicate identical instances of the minority class
Establish a balanced class distribution
Random Oversampling
No responseSubscribed to magazine
Benefits
• Helps detect rare target levels
• No loss of information
Risks
• Biases predictions (correctable)
• Increases prediction variability
• Increases processing time
1 … Number of subscriptions … Many
Few
… D
ays s
ince last
purc
hase
… M
any
![Page 128: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/128.jpg)
rather some case studies ...!
Ready for more theory…?
x
![Page 129: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/129.jpg)
• Sampling issues in Data Mining
• Case study 1: Direct Marketing• Cross-selling of Magazine subscriptions
• Effect of data preprocessing: Sampling
• Interaction of Sampling with Scaling & Coding
• Case study 2: Credit & Behavioral Scoring• Predicting consumer credit default
• Effects of sample size
• Effects of sample distribution
• Case study 3: Online Shopping Behaviour• Predicting consumer shopping channel choice
• Sample distribution & multiple classes
• Conclusion & Take-aways
Agenda
![Page 130: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/130.jpg)
• Sell a magazine subscription to existing customers
• Whom to send mail to? (Which customers are most likely to respond?)
• How many customers to contact? (What is the optimal mailing size?)
Corporate project with leading German Publishing HouseProvided data set of past mailing campaigns
Benchmark novel methods against in-house SPSS Clementine
Explore Neural Networks (NN) an Support Vector Machines (SVM)
Business Case:
Direct Marketing/Response Optimization
![Page 131: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/131.jpg)
Smaller mailing (number of letters sent) lower costs (Euro 1.- per letter)
Higher response rate higher revenue
More specific mailing lower cost
More relevant information higher customer satisfaction
Benefits of Direct Marketing
Simple With data mining
Addressees 100.000 Top 40% = 40.000
Cost 2€/mail = 200.000€ 2,5€/mail = 100.000€
Response rate 0,5% = 500 1,0% = 400
Sales volume 300€ 300€
Sales volume 150.000€ 120.000€
Revenue -50.000€ 20.000€
![Page 132: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/132.jpg)
NN get worse with learning …
%Pred. C 0
Pred. C 1
Sum
C 0 72.96 27.04 100
C 1 62.02 37.98 100
134.98 65.02 55.47
%Pred. C 0
Pred. C 1
Sum
C 0 52.87 43.37 100
C 1 47.13 56.63 100
100 100 54.75
%PredC 0
PredC 1
Sum
C 0 61.86 38.14 100
C 1 55.09 44.81 100
116.95 82.95 54.26
• Wish to implement Neural Networks for next campaign• In-house team (with no NN knowledge) outperformed us EVERY TIME!• Analyzed software, training parameters, etc. internal competition• Observed expert in building models … !
![Page 133: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/133.jpg)
Scale numerical
features
Adjust imbalanced
class distributions
Decide on sample
size and method
Experimental Design:
Different data pre-processing
Handle categorical
features
Select useful
features
Handle outliers
Different SamplingOver-& Undersampling
Different Encodingn, n-1, thermo, ordinal
Different ScalingDiscretise, Standardise
Evaluate across 3 algorithms:
Neural Networks (MLPs), Support Vector Machines & Decision Trees
![Page 134: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/134.jpg)
Multifactorial design to evaluate impact across multiple methods
Neural Networks (NN)
Support Vector Machines (SVM)
Decision Trees (CART)
Dataset Structure
Data set size• 300,000 customer records• 4,019 subscriptions sold• Response rate of 1.3%
Data set structure• 18 categorical features• 35 numerical features• Binary target variable
Evaluated the Impact of Data Preprocessing
• Data Sampling (over sampling vs. undersampling)
• Categorical attribute Encoding (N, N-1, thermo, ordinal)
• Continuous attribute Projection (Binning vs. Normalisation)
• Continuous attribute Scaling ( [0,+1] vs. [-1,+1] range)
![Page 135: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/135.jpg)
Different balancing in the training data
Original distribution in the test data (65,000 instances)
Sampling
Data partition (number of records)
Oversampling Undersampling
Data subset Class 1 Class -1 Class 1 Class -1
Training set 20,000 20,000 2,072 2,072
Validation set 10,000 10,000 1,035 1,035
SUM 30,000 30,000 3,107 3,107
Test (hold-out) set 912 64,088 912 64,088
Created 2 Dataset Sampling candidates
![Page 136: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/136.jpg)
Oversampling outperforms undersampling consistently!
Gain in Lift depends on method (different sensitivity)
Oversampling has higher impact than data coding & scaling
Results
Increase
Increase
Increase
![Page 137: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/137.jpg)
Binning & Scaling of continuous attributes irrelevant for all methods!
Use Undersampling & N-1 encoding with SVM & NN
Best preprocessed SVM lift of 0.645 on test set … BUT …
Recommendations from Case Study
• Sampling • Oversampling outperfoms undersampling for all methods
• Undersampling: better in-sample results & worse out of sample
• Choice of method • NN & SVM better than CART
• Encoding & Projection• SVM: avoid Ordinal coding (e.g. 1,2,3) all other similar (incl. N !)
• NN: avoid standardization & ordinal encoding
• DT / CART: use temperature, all others similar (incl. ordinal)
![Page 138: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/138.jpg)
Results are consistent across error measures
Experiments allow identification of „best practices‟ to model methods
Best-practice preprocessing varies between methods
Results across Pre-processing
Preprocessing: higher impact than method selection Lift-variation per method from Sampling/Scaling/Coding
> Difference of Lift between competing methods!
DTSVMNN
Method
0,65
0,64
0,63
0,62
0,61
0,60
Lif
t te
st
Lift performance onTest data subset
DTSVMNN
Method
0,58
0,57
0,56
0,55
0,54
0,53
AM
te
st
Arithmetic Mean Performanceon Test data subset
DTSVMNN
Method
0,58
0,57
0,56
0,55
0,54
0,53
0,52
0,51
0,50
GM
te
st
Geometric Mean Performanceon Test data subset
DPP causes 50%-70% of the
differences between models
![Page 139: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/139.jpg)
• Sampling issues in Data Mining
• Case study 1: Direct Marketing• Cross-selling of Magazine subscriptions
• Effect of data preprocessing: Sampling
• Interaction of Sampling with Scaling & Coding
• Case study 2: Credit & Behavioral Scoring• Predicting consumer credit default
• Effects of sample size
• Effects of sample distribution
• Case study 3: Online Shopping Behaviour• Predicting consumer shopping channel choice
• Sample distribution & multiple classes
• Conclusion & Take-aways
Agenda
![Page 140: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/140.jpg)
Business Case: Predicting
Customer Online Shopping Adoption
• Traditional buying process is offline & simultaneous “bricks” store
• Introduction of the Internet changes consumer behaviour• Seek information online & offline
• Purchasing online & offline
Changing purchasing behaviour through internet adoption
Changing purchasing behaviour through Technology Acceptance
• Development of heterogeneous Purchasing Behaviour• Example: Purchasing electronic durable consumer goods
• Search for product info (e.g. video cameras) online
test product in-store
search for best deal on internet & purchase
Search for Information Online Purchase Online
Search for Information Offline Purchase Offline
Online
Shoppers
Non-Internet
Shoppers
Browsers
![Page 141: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/141.jpg)
Stages of Internet Adoption
1. OFFLINE BUYERS
Information gathering
& purchasing in Stores
2.BROWSERS
Information gathering online
& purchasing in stores
3.ONLINE BUYERS
Information gathering
& purchasing online
![Page 142: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/142.jpg)
Motivation
DIDIER: Marketing Modelling
• Econometric / Marketing Domain
• Seeks to explain how customers behave in
online shopping
• Use of „black-box” logistic regression
models
Models class membership to identify
causal variables that explain choices
Descriptive & Normative Modelling
SVEN: Data Mining Perspective
• IS/OR/MS Domain Data Mining
• Seeks to accurately predict regardless of
explanation why customers buy
• Use of “black-box” methods from
computational intelligence
Models class membership to
accurately classify unseen instances
Predictive Modelling
same dataset & same objectives & similar methods
Conflicting “best practice” approaches to modelling
Outside of most software simulators!!! Implicit knowledge?
… WHO IS “CORRECT”? WHAT IS THE IMPACT?
Best practices
balance datasets for distribution
representative of population
Use ordinal variables & nominal variables
without recoding
Do not normalise / scale data
Best practices
Rebalance datasets for equal distribution
of target variables
Recode ordinal binary scale
Rescale & normalise data to facilitate
learning speed etc.
![Page 143: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/143.jpg)
Dataset
• Survey on Internet Shopping Behaviour• 5500 UK households 685 respondents
• Adjusted for age, income etc. of customers (older less likely to buy)
• Adjusted for product specific risk of online shopping for branded durable consumer goods (inspection required to some extent)
• 73 questions on factors related to internet shopping, products etc.
Models Output VariablesInput Variables
Demographics
Internet
specific
Factors
Online
shopping
specific
Factors
Logistic Regression
Neural Networks
Class 1:
Browse Ônline &
Buy Online
Class 2:
Browse Online &
Buy Offline
Class 3:
Browse &
Buy Offline
Online Shopping Factors:
“Going to the shops is as convenient
as Internet shopping”
“I would buy online if products are
branded” etc. [1=strongly agree; …]
Demographic Factors
Age, Gender, Income
Internet Utility Factors
Score from 6 correlated variables
Mixed scale of nominal, ordinal, interval
![Page 144: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/144.jpg)
Imbalanced Classification problem
UndersamplingOversamplingImbalanced
Dataset
Offline-Shoppers
BrowsersOnline-Shoppers
Offline-Shoppers
BrowsersOnline-Shoppers
Offline-Shoppers
BrowsersOnline-Shoppers
400
300
200
100
0
Co
un
t
Test
Validation
Training
Data Subset
UndersamplingOversamplingImbalanced
Dataset
Offline-Shoppers
BrowsersOnline-Shoppers
Offline-Shoppers
BrowsersOnline-Shoppers
Offline-Shoppers
BrowsersOnline-Shoppers
400
300
200
100
0
Co
un
t
Test
Validation
Training
Data Subset
• Split of Dataset for Training, Validation and Test {50%;25;25%}• Distribution of target classes is skewed
{65% online buyers; 22.5% browsers; 12.5% offline shoppers}• Rebalancing of data sets through over- & undersampling)
![Page 145: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/145.jpg)
Results without DiscretisationLogist.Reg. True
Value
Training Data Test Data
Dataset Online Browse Offline Online Browse Offline
Original Online 93.36 5.17 1.48 88.89 7.78 3.33
Imbalanced Browser 62.77 23.40 13.83 49.39 22.58 29.03
Offline 36.54 17.31 46.15 35.29 29.41 35.29
Under- Online 57.69 30.77 11.54 64.44 23.33 12.22
Sampling Browser 26.92 48.08 25.00 32.26 25.81 41.94
Offline 17.31 21.15 61.54 29.41 35.29 35.29
Over- Online 68.27 24.35 7.38 74.44 16.67 8.89
Sampling Browser 30.63 43.91 25.46 35.48 29.03 35.48
Offline 16.97 19.93 63.10 29.41 29.41 41.18
Neural Net Training Data Test Data
Dataset Online Browse Offline Online Browse Offline
Original Online 86.19 12.71 1.10 86.67 8.89 4.44
Imbalanced Browser 53.13 31.25 15.63 41.94 35.48 22.58
Offline 25.17 28.57 45.71 29.41 35.29 35.29
Under- Online 44.86 40.00 17.14 27.78 58.89 13.33
Sampling Browser 14.29 48.57 37.14 16.13 32.26 51.61
Offline 8.57 20.00 71.43 11.76 41.18 47.06
Over- Online 81.22 18.23 0.55 61.11 22.22 16.67
Sampling Browser 14.92 83.43 1.66 19.35 77.42 3.23
Offline 15.52 0.55 99.45 0.00 11.76 88.24
MCRtrain=54.3%
MCRtest =48.9%
MCRtrain=55.8%
MCRtest =41.8%
MCRtrain=58.4%
MCRtest =48.2%
MCRtrain=54.4%
MCRtest =52.5%
MCRtrain=54.9%
MCRtest =35.7%
MCRtrain=88.0%
MCRtest =75.6%
Mean Classification Rate (%)
![Page 146: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/146.jpg)
Results with Discretisation of OrdinalLogist.Reg. True
Value
Training Data Test Data
Dataset Online Browse Offline Online Browse Offline
Original Online 91.51 6.64 1.85 85.56 7.78 6.67
Imbalanced Browser 54.26 36.17 9.57 48.39 32.26 19.35
Offline 26.92 17.31 55.77 58.82 47.62 17.65
Under- Online 71.15 21.15 7.69 55.56 24.44 20.00
Sampling Browser 17.31 65.38 17.31 67.74 6.45 25.81
Offline 15.38 11.54 73.08 58.82 0.00 41.18
Over- Online 68.63 22.88 8.49 70.0 21.11 8.89
Sampling Browser 17.34 56.83 25.83 12.90 58.06 29.03
Offline 13.28 14.02 72.69 17.65 23.53 58.82
Neural Net Training Data Test Data
Dataset Online Browse Offline Online Browse Offline
Original Online 96.13 3.87 0.00 84.44 11.11 4.44
Imbalanced Browser 68.75 28.13 3.13 64.52 22.58 12.90
Offline 40.00 14.29 45.17 58.82 11.76 29.41
Under- Online 57.14 40.00 2.86 25.56 72.22 2.22
Sampling Browser 34.29 54.29 11.43 67.74 29.03 3.23
Offline 14.29 31.43 54.29 52.94 17.65 29.41
Over- Online 98.34 1.10 0.55 58.89 24.44 16.67
Sampling Browser 0.00 100.0 0.00 3.23 83.87 12.90
Offline 0.00 0.00 100.0 0.00 5.88 94.12
MCRtrain=61.15%
MCRtest =45.1%
MCRtrain=69.9%
MCRtest =34.4%
MCRtrain=66.0%
MCRtest =62.3%
MCRtrain=56.5%
MCRtest =45.5%
MCRtrain=55.2%
MCRtest =28.0%
MCRtrain=99.5%
MCRtest =79.0%
Mean Classification Rate (%)
![Page 147: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/147.jpg)
Oversampling outperforms other samplings- Across Different Datasets
- Across various data preprocessing
Methods show different sensitivity to Sampling- More variation from sampling, coding & scaling than between methods
- Using different preprocessing variants is important in modeling
Various sophisticated extensions exist- SMOTE (Synthetic Minority Oversampling Technique)
- K-nearest Neighbor sampling (removal / creation)
- One-class learning etc. …
Extend your bad of tricks …- … and experiment with imbalanced sampling!
Summary
![Page 148: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/148.jpg)
Sven F. CroneLancaster University Management School
Centre for Forecasting
Lancaster, LA1 4YX
email [email protected]
Questions?
1 1(1 ) t t tSY Y SY
![Page 149: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/149.jpg)
Exploring Innovation
“Online panel” vs “Online „streaming/convenience”
sampling
![Page 150: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/150.jpg)
Unfortunately, the presenters of iVOX &
Corelio cannot share their presentation
with the BAQMaR community due to
reasons of confidentiality!
![Page 151: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/151.jpg)
![Page 152: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/152.jpg)
![Page 154: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/154.jpg)
![Page 155: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/155.jpg)
![Page 156: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/156.jpg)
![Page 157: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/157.jpg)
![Page 158: BAQMaR 2008](https://reader033.vdocuments.net/reader033/viewer/2022052820/545c0ed3af7959c8098b4622/html5/thumbnails/158.jpg)