Event: Dataconomy Düsseldorf, 19.10.2017
Speaker: Dr. Christoph Tempich
Weitere Tech-Vorträge: https://www.inovex.de/de/content-pool/vortraege/
Tech-Artikel im inovex-Blog: https://www.inovex.de/blog/
WSO2Con2024 - Enabling Transactional System's Exponential Growth With Simplicity
Data Product Discovery: The product perspective on digital transformation
1. Data product discovery
The product perspective on digital
transformation
Dr. Christoph Tempich
Chief Data Economist
Köln, www.inovex.de 19.10.2017
#Datenprodukte #dataproducts @ctempich
3. 3G Linden, B Smith, J York: Amazon. com recommendations: Item-to-item collaborative filtering - IEEE Int. Comp., 2003
How did they calculate them at the beginning?
2003 2017
Item-to-item
collaborative filtering
Hybrid recommenders
Deep learning
Random
4. 4Quelle: inovex Case Study Recommendations bei mobile.de
Recommendations @ inovex
Germany’s biggest vehicle market place
User Benefits
› engagement
› inspiration
› relevance
Business Benefits
› high click-through-rate
› small exit- & bounce-rates
6. 7
Data products: types
Data as a Service
Data-enhanced
Products
Data as Insights
Type 1 Type 2 Type 3
› Autonomous driving
› Recommendations
› Weather data › Marketing planning
#Datenprodukte @ctempich @thomasleiterman
11. 12Von Watzmann - Eigenes Werk, Gemeinfrei, https://commons.wikimedia.org/w/index.php?curid=8391805
Variance = Uncertainty
A measure for the value of information
12. 13
Stakeholder uncertainty
What is the problem and can data solve it?
Stakeholder:
Person who runs
• Increased
stamina
• Weight mgmt.
• Increased
performance
• Fun
• Continuous
improvement
• Motivating • Supports
performance
gains
• Supports
performance
gains
• Heart attack • Broken joints • No improvement
• Too exhausting
• Demotivating • Demotivating • No training success
Customer Jobs
Deciding to run
Selecting a
runner
Workout plan
Course
selection
Run Recover
Best
case
Worst
case
14. 15Quelle: Laura Dorfer: Datenzentrische Geschäftsmodelle als neuer Geschäftsmodelltypus …, 2016.
Value proposition and readiness to pay
Data provider and consumer may differ!
User Buyer
Social interaction
Entertainment
Curiosity
Decision support
Transparency
Information
Readiness to payValue proposition
15. 16
Looking into the future
Example: Training support
• Social interaction
• Which course have I taken?
• Information
• How fast am I running currently?
• Decision
• When should I run again?
BasedontransactiondataBasedon
masterdata
Value Proposition Service
• Information
• Shoes
Future
Present
Past
Data item
Information about
18. 19
Recognition of house numbers for
Google Street View,
Feedback loop: Which house number is this?
by means of involving users in
recognizing the correct number.
Bsp.: https://www.google.com/recaptcha/intro/android.html
Quelle: http://stadt-bremerhaven.de/google-streetview-und-
captchas-beide-profitieren-voneinander/
19. 20Quelle: Data Value Matrix
Data Value Matrix
Improves
same
service
Improves
other
service
Machine learning
Aggregation
Tesla fleet learning
Improves autonomous driving Mobile.de User Profile
Show appropriate ads
Google reCAPTCHA
Improves Google Street View
STRAVA
Training course highlighting
nike plus
Music selection
Mobile.de Recommendation
Recommend similar cars
21. 1. Start with the customer!
2. What is the problem and can data solve it?
3. Test your hypothesis using the simplest algorithm
4. Data provider and consumer often differ
5. Close the feedback loop
22#Datenprodukte #dataproducts @ctempich