Customer Data
at De Persgroep
Luc Verbist CIO De Persgroep 10 – 10 – 2013
De Persgroep
Marketing B2C
Strategy
Architecture
Privacy
Agenda
NEWSPAPERS
Activities
N°
1
Activities
MAGAZINES
N°
1
Activities
Print products
Activities
RADIO
Activities
TELEVISION
Activities
DIGITAL
Classified
Sites
E-commerce
iPad &
mobile
News
Pay-per-view
Activities
MOBILE TELEPHONY
Cellphone
Operator
Financial Results
(Mio €)
De Persgroep
Marketing B2C
Strategy
Architecture
Privacy
Agenda
Marketing B2C
As Is
Clustering Segmentation
James Bond Movies 65.000 in 5 weeks Cooking Books 135.000 on 1 day DVD players 35.000 on 1 day
Reactions not logged Operational processing
Social Media not followed No logging
Surfers or behaviour Behavior on App
not logged
To Be
Mass Marketing + Micro/Individual Marketing
To Be
1.200.000 Customers 11.000.000 Prospects 1.400.000 Deliveries/day 10.000 Calls/day 11.000.000 Prospects 2.000.000 Marketing deliveries/year 10.000 Tweets/day titels 100.000 Tweets/day customers 120.000.000 Pageviews/dayMarketing B2C
De Persgroep
Marketing B2C
Strategy
Architecture
Privacy
Agenda
CRM strategy for DIGITAL (focus on traffic building) CRM strategy for PRINT (focus on recruitment) Little synergy
1. Integrated strategy for print and digital on the basis of a uniform view of the consumer in all its relationships with a title.
2. To map Recency and Frequency of reading habits in each channel.
3. Infer / clustering interest according to preferences for different categories (region, sports, politics…)
Opportunities :
Readers Marketing:
Readers recruitment
Upsell and cross-Sales
Conservation
Reduce churn
Advertising use:
More targetted campaigns
Personalize ads
Multi channel campaings
De Persgroep
Marketing B2C
Strategy
Architecture
Privacy
Agenda
Architecture
Challenges :
Multiple data sources
Multiple formats
No or different Unique Identifiers
Speed of operation is essential
Huge volumes of data
Manage Operational cost
Moving data requires (some) technical expertise
No real standards defined
Architecture
Active Newspaper Subscriptions Online Customers & Prospects Newspaper Subscription & Payment History Customer & Prospects Contacts Integrati on Data Custo mer Consumer Behaviour Interests aggregates External Data Data Cleansing, Quality check, Enrichment Subscriber Communication History Digital Newspaper downloads Website visits Newsletter subscriptionsArchitecture
Business rules :
How to identify/recognize a unique human being
Matching algoritm Windowkey Pattern Co m pa ny N am e G ender Initi al s Fi rs t N am e Las t N am e Str ee t H o us e N r B us N r Zi p C d Lo cat io n Em ai l Bi rth Dt Tel N r G sm N r V at N r A bo n Id NAW
100/101 First + Last + Address + Email E E E E E E E
First + Last + Address + Email F F F E E E E
110/111 First + Last + Address E E E E E E
First + Last + Address F F F E E E
120/121 Initial + Last + Address + Birthdate E E E E E E E
Initial + Last + Address + Birthdate E F F E E E E
130/131 Gender + Initial + Last + Address E E E E E E E
Gender + Initial + Last + Address E E F F E E E
140 Email + First + Last E E E
Email + First + Last F F E
150 Email + Initial + Last E F E
160 Email + Birtdate E E 170 Email + Abon Id E E 180 Email + Gsm Nr E E Gsm Nr 190 Gsm Nr + First + Last E E E Gsm Nr + First + Last F F E 200 Gsm Nr + Last + Initial E E 210 Gsm Nr + Abon Id E E Tel Nr 220
Tel Nr + First + Last E E E Tel Nr + First + Last F F E
Architecture
Business rules :
Assign quality levels to data
Use specialised software or services (e.g.Trillium)
Use exact and fuzzy matching algoritm
What level of correctness is required? (Pareto)
Architecture
A H G B C F E D ...Federation of companies : enriching & correcting data Starting from the data strenght of each company
Strategy
Deduplication and identification a real tough job.
Handling massive volumes is something your staff needs to learn.
Involve experts (from outside your company).
Nothing is perfect yet in de Big Data market.
Hard to justify the project as you can not predict tangible benefits upfront.
De Persgroep
Marketing B2C
Strategy
Architecture
Privacy
Agenda
Privacy
Quid Privacy legislation?
Personal data = any information of any type relating to an
identified or identifiable natural person.
Data must be processed fairly and lawfully and be up to date
Privacy
Quid Privacy legislation?
Purpose Limitation : Specific, explicit and legitimate
Formalities Prior notification Belgian Commission for the
Protection of Privacy
Privacy
What is the risk?
Reputation
Future data protection rules in EU : 2% of annual ww turnover
Sanctions : administrative, criminal, civil and damage
Privacy
Quid Privacy legislation?
DP wants * highest possible transparency to consumer and his data * possibility of control and adjustment to own data
Project My Account
Privacy
Quid Privacy legislation?
Non-intrusive questions or polls reveal more about a person than he/she realizes. What kind of swim shorts do you prefer?
Your choice reveals a lot about your age.
Q&A
Luc Verbist CIO De Persgroep [email protected]