3. Agenda
• Current Stat. of (Big) Data Science
• KFF of building a data science team
– Add data scientists
– Analysis with reports than models
– Political conflicts
– Sales judgment
• Some directions
4. Current Stat. of (Big) Data Science
Source: Capgemini Consulting, “Big Data Survey”, November 2014.
9. KFF1: Add Data Scientists
BDSales + AS
Sales + CM
Data BD
Data Engineer +
Data Scientist
Conversions +
3rd
Tracking
10. KFF2: Analysis with reports
than models
資料量大 資料多樣性
資料輸入
和處理速度快
資料真實性
Challenges of Big Data - 4V
11. KFF2: Analysis with reports
than models
Russ Merz, An Integrated Model of Media Satisfaction and Engagement: Theory, Empirical
Assessment and Managerial Implications, Journal of Applied Marketing Theory, Nov 2011
BIG DATA
Hypotheses
Machine Learning
Data Mining
Machine-generated
All, Hyper space, …
Volume, Velocity, Variety, Veracity
deductive inductive
Cases
Models
Models
Cases
12. KFF2: Analysis with reports
than models
Segments Reports
For Human
(Explanatory)
Models Data-driven
Actions
Efficiency Intelligence Effectiveness
Data Science is the art of turning data into actions.
13. KFF3: Political Conflicts
Data Science Venn Diagram
Cross-functional data
Performance attribution
Analysts
Legacy integration
CTO /
Tech lead
14. KFF4: Sales Judgment
• Data late effect
– After a data management system
– After data links and accumulation
– After experiments & optimization
• Pricing by CPM/CPC vs CPI/CPS
• Career path of sales head
• Positioning & Orientation
– Product companies: Apple, Google, AppNexus
– Marketing companies: Microsoft, Uber
– Sales companies: Oracle, SAP…
15. Some Directions
• Planned with Fail Fast
• Full functions with a Data Lab
– Develop operational data systems first?
• Secure funding
• Strong talents/Trust
– Experimenter with
Stat/MLDM
• A strong leader
Machine learning:
Building models, Validating models, Scoring the data
Why machine learning is not enough:
Lack of perspective, credibility, complexity
David M S mith
VP Marketing and C ommunity
R evolution Analytic s
Cost-per-click pricing optimized against cost-of-sale
Post-click sales attribution as the measure of effectiveness