1. From data-driven startup to large company in a decade
Dr. Andreas Both
Head of Research and Development
Unister GmbH
Germany
European Data Forum 2012
Copenhagen, June 6-7, 2012

2. Claim
. . . challenges of Big Data and the emerging Data Economy
and to develop suitable action plans for addressing these
challenges . . .
Slide 2 Dr. Andreas Both, Head of R&D, Unister

3. Claim
. . . challenges of Big Data and the emerging Data Economy
and to develop suitable action plans for addressing these
challenges . . .
Slide 2 Dr. Andreas Both, Head of R&D, Unister

4. Claim
. . . challenges of Big Data and the emerging Data Economy
and to develop suitable action plans for addressing these
challenges . . .
Slide 2 Dr. Andreas Both, Head of R&D, Unister

5. Claim
. . . challenges of Big Data and the emerging Data Economy
and to develop suitable action plans for addressing these
challenges . . .
Slide 2 Dr. Andreas Both, Head of R&D, Unister

6. Personal Retrospective . . .
Slide 3 Dr. Andreas Both, Head of R&D, Unister

7. Personal Retrospective . . .
. . . joining Unister, April 2010
eCommerce is challenging
complexity of business
everything has to work smoothly
punishment comes quickly
Slide 3 Dr. Andreas Both, Head of R&D, Unister

8. Personal Retrospective . . .
. . . joining Unister, April 2010 some lessons
eCommerce is challenging improve your business model
complexity of business care about our customers
everything has to work smoothly
punishment comes quickly
Slide 3 Dr. Andreas Both, Head of R&D, Unister

9. Personal Retrospective . . .
. . . joining Unister, April 2010 some lessons
eCommerce is challenging improve your business model
complexity of business care about our customers
everything has to work smoothly rapide impact of decisions
punishment comes quickly increasing data complexity
Slide 3 Dr. Andreas Both, Head of R&D, Unister

10. It’s about the business activities, stupid.
Should we really talk about data?
Why should we talk about data processes?
Why should we talk about the data analytics?
Slide 4 Dr. Andreas Both, Head of R&D, Unister

11. It’s about the business activities, stupid.
Should we really talk about data?
Why should we talk about data processes?
Why should we talk about the data analytics?
Is it just about the business?
Is data a core value?
Slide 4 Dr. Andreas Both, Head of R&D, Unister

12. IT companies vs. traditional business
classic business e-business
need natural ressource need data
to establish processes to establish processes
depending on location independent from location
Slide 5 Dr. Andreas Both, Head of R&D, Unister

13. IT companies vs. traditional business
classic business e-business
need natural ressource need data
to establish processes to establish processes
depending on location independent from location
SME will find many niches within the market.
Slide 5 Dr. Andreas Both, Head of R&D, Unister

14. IT companies vs. traditional business
classic business e-business
need natural ressource need data
to establish processes to establish processes
depending on location independent from location
SME will find many niches within the market.
. . . establish large companies too!
Slide 5 Dr. Andreas Both, Head of R&D, Unister

15. Unister’s Success Story
Internet startup mainly located in Leipzig, Germany
private limited company (German GmbH)
founded in 2002, 5 founders
managing director: Thomas Wagner
a
Slide 6 Dr. Andreas Both, Head of R&D, Unister

16. Unister’s Success Story
Internet startup mainly located in Leipzig, Germany
private limited company (German GmbH)
founded in 2002, 5 founders
managing director: Thomas Wagner
eCommerce company, B2C
national and international activities
> 40 web-portals and services
> 13.22 mio. unique user / month in Germanya
a
AGOF e.V. / internet facts 2012-01
Slide 6 Dr. Andreas Both, Head of R&D, Unister

17. Unister’s Success Story
Internet startup mainly located in Leipzig, Germany
private limited company (German GmbH)
founded in 2002, 5 founders
managing director: Thomas Wagner
eCommerce company, B2C
national and international activities
> 40 web-portals and services
> 13.22 mio. unique user / month in Germanya
IT-driven approach
a
AGOF e.V. / internet facts 2012-01
Slide 6 Dr. Andreas Both, Head of R&D, Unister

18. Unister’s Success Story
Slide 7 Dr. Andreas Both, Head of R&D, Unister

19. Unister’s Success Story
Number of Employees
1530
1157
Employees
701
372
185
106
7 38
1
2003 2004 2005 2006 2007 2008 2009 2010 2011
Slide 7 Dr. Andreas Both, Head of R&D, Unister

20. IT companies
Data is the foundation
getting data is tough
having data is not enough
managing data is challenging
Slide 8 Dr. Andreas Both, Head of R&D, Unister

21. IT companies
Data is the foundation
getting data is tough
having data is not enough
managing data is challenging
Next parts of the talk
Data Access
Data Integration
(Big) Data Analyses
Slide 8 Dr. Andreas Both, Head of R&D, Unister

22. IT companies
Data is the foundation
getting data is tough
having data is not enough
managing data is challenging
Next parts of the talk
Data Access
← Steps a start-up has to tackle!
Data Integration
(Big) Data Analyses
Slide 8 Dr. Andreas Both, Head of R&D, Unister

23. IT companies
Data is the foundation
getting data is tough
having data is not enough
managing data is challenging
Next parts of the talk
Data Access
← Steps a start-up has to tackle!
Data Integration
Steps Unister had tackled.
(Big) Data Analyses
Slide 8 Dr. Andreas Both, Head of R&D, Unister

24. IT companies
Data is the foundation
getting data is tough
having data is not enough
managing data is challenging
Next parts of the talk
Data Access
← Steps a start-up has to tackle!
Data Integration
Steps Unister had tackled.
(Big) Data Analyses Steps Unister has to tackle.
Slide 8 Dr. Andreas Both, Head of R&D, Unister

25. Data Access
Slide 9 Dr. Andreas Both, Head of R&D, Unister

26. Data Access
Observations
business models need access
of data
support, description,
enrichment, . . .
Slide 10 Dr. Andreas Both, Head of R&D, Unister

27. Data Access
Observations Unister’s Success (step 1)
business models need access was capable of integrating
of data many data sets
support, description,
user-focussed data
enrichment, . . .
Slide 10 Dr. Andreas Both, Head of R&D, Unister

28. Data Access
Observations Unister’s Success (step 1)
business models need access was capable of integrating
of data many data sets
support, description,
user-focussed data
enrichment, . . .
eCommerce demand
local information
link to local events
(legal) contraints
→ such data not available
Slide 10 Dr. Andreas Both, Head of R&D, Unister

29. Data Access
Observations Unister’s Success (step 1)
business models need access was capable of integrating
of data many data sets
support, description,
user-focussed data
enrichment, . . .
Challenges eCommerce demand
Open Data should be local information
established link to local events
Standards have to be defined (legal) contraints
and followed
→ such data not available
Slide 10 Dr. Andreas Both, Head of R&D, Unister

30. Data Access: Needed Actions
Open Data initiatives need support
enhanced tool support
political commitment
Slide 11 Dr. Andreas Both, Head of R&D, Unister

31. Data Access: Needed Actions
Open Data initiatives need support
enhanced tool support
political commitment
will establish (local) data economics
locally connected companies could grow
Slide 11 Dr. Andreas Both, Head of R&D, Unister

32. Data Integration
Slide 12 Dr. Andreas Both, Head of R&D, Unister

33. Data Integration
Observations
bread and butter of
data-driven companies
needs much effort
Slide 13 Dr. Andreas Both, Head of R&D, Unister

34. Data Integration
Observations Unister’s Success (step 2)
bread and butter of fusion of different data sets
data-driven companies leads to good user experience
needs much effort
Slide 13 Dr. Andreas Both, Head of R&D, Unister

35. Data Integration
Observations Unister’s Success (step 2)
bread and butter of fusion of different data sets
data-driven companies leads to good user experience
needs much effort
eCommerce demand
matching processes
integration tools
Slide 13 Dr. Andreas Both, Head of R&D, Unister

36. Data Integration
Observations Unister’s Success (step 2)
bread and butter of fusion of different data sets
data-driven companies leads to good user experience
needs much effort
Challenges eCommerce demand
establish distributed matching processes
knowledge base
integration tools
Linked Data paradigm
NOSQL + SQL
Slide 13 Dr. Andreas Both, Head of R&D, Unister

37. Data Integration: Needed Actions
support Linked Data Cloud
companies need sound and solid data sets
Slide 14 Dr. Andreas Both, Head of R&D, Unister

38. Data Integration: Needed Actions
support Linked Data Cloud
companies need sound and solid data sets
research on scalable data integration
processes
Cloud Computing → Big Data challenge
Slide 14 Dr. Andreas Both, Head of R&D, Unister

39. Data Analyses and Big Data
Slide 15 Dr. Andreas Both, Head of R&D, Unister

40. Data Analyses and Big Data
Observations
disseminate, understand and
ultimately benefit from
increasing volumes of data
example: social networks
Slide 16 Dr. Andreas Both, Head of R&D, Unister

41. Data Analyses and Big Data
Observations Unister’s Success (step 3)
disseminate, understand and defining data analysis
ultimately benefit from processes with impact
increasing volumes of data pareto-optimal processes lead
example: social networks to good coverage
analysis came to a limit
because of many segments
Slide 16 Dr. Andreas Both, Head of R&D, Unister

42. Data Analyses and Big Data
Observations Unister’s Success (step 3)
disseminate, understand and defining data analysis
ultimately benefit from processes with impact
increasing volumes of data pareto-optimal processes lead
example: social networks to good coverage
analysis came to a limit
because of many segments
eCommerce demand
descriptive analyses processes
higher-level process interfaces
good developers
Slide 16 Dr. Andreas Both, Head of R&D, Unister

43. Data Analyses and Big Data
Observations Unister’s Success (step 3)
disseminate, understand and defining data analysis
ultimately benefit from processes with impact
increasing volumes of data pareto-optimal processes lead
example: social networks to good coverage
analysis came to a limit
because of many segments
Challenges eCommerce demand
... descriptive analyses processes
higher-level process interfaces
good developers
Slide 16 Dr. Andreas Both, Head of R&D, Unister

44. Data Analyses and Big Data: Global Movement
source: ucsd.edu
Slide 17 Dr. Andreas Both, Head of R&D, Unister

45. Data Analyses and Big Data: Challenges
source: hadapt.com
Slide 18 Dr. Andreas Both, Head of R&D, Unister

46. Data Analyses and Big Data: Challenges
The 3 V
Volume
Varity
Velocity
Slide 19 Dr. Andreas Both, Head of R&D, Unister

47. Data Analyses and Big Data: Challenges
The 3 V
Volume
Varity
Velocity
The +2 V
Virality
Viscosity
Slide 19 Dr. Andreas Both, Head of R&D, Unister

48. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
Slide 20 Dr. Andreas Both, Head of R&D, Unister

49. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
Slide 20 Dr. Andreas Both, Head of R&D, Unister

50. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
Slide 20 Dr. Andreas Both, Head of R&D, Unister

51. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
good support: Cloud Computing, . . .
Slide 20 Dr. Andreas Both, Head of R&D, Unister

52. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
good support: Cloud Computing, . . .
need people to operate on the systems
Slide 20 Dr. Andreas Both, Head of R&D, Unister

53. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
good support: Cloud Computing, . . .
need people to operate on the systems
→ ok: some experience available
Slide 20 Dr. Andreas Both, Head of R&D, Unister

54. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
good support: Cloud Computing, . . .
need people to operate on the systems
→ ok: some experience available
need people to develop applications
Slide 20 Dr. Andreas Both, Head of R&D, Unister

55. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
good support: Cloud Computing, . . .
need people to operate on the systems
→ ok: some experience available
need people to develop applications
→ bad: rarely teached
Slide 20 Dr. Andreas Both, Head of R&D, Unister

56. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
good support: Cloud Computing, . . .
need people to operate on the systems
→ ok: some experience available
need people to develop applications
→ bad: rarely teached
need people to think about using the network effect
Slide 20 Dr. Andreas Both, Head of R&D, Unister

57. Data Analyses and Big Data: Needed Actions
handle this is all about knowledge . . .
levels of challenge
need systems for big data analyses
good support: Cloud Computing, . . .
need people to operate on the systems
→ ok: some experience available
need people to develop applications
→ bad: rarely teached
need people to think about using the network effect
→ very bad: Talent Gap
Slide 20 Dr. Andreas Both, Head of R&D, Unister

58. Data Analyses and Big Data: Potential
Big data – The next frontier for innovation, competition,
and productivity
(McKinsey May 2011)
Slide 21 Dr. Andreas Both, Head of R&D, Unister

59. Data Analyses and Big Data: Potential
Big data – The next frontier for innovation, competition,
and productivity
(McKinsey May 2011)
Plenty of possibilities!
Slide 21 Dr. Andreas Both, Head of R&D, Unister

60. Summary: Most important activities
Open Data will give a push
well-developed tools are crucial for SME
talent gap has to be tackled
Slide 22 Dr. Andreas Both, Head of R&D, Unister

61. Conclusion
Data Access, Data Integration, Data Analyses
Slide 23 Dr. Andreas Both, Head of R&D, Unister

62. Conclusion
Data Access, Data Integration, Data Analyses
Big Data
Slide 23 Dr. Andreas Both, Head of R&D, Unister

63. Conclusion
Data Access, Data Integration, Data Analyses
Big Data successful companies
Slide 23 Dr. Andreas Both, Head of R&D, Unister

64. From data-driven
startup to large company
in a decade
Dr. Andreas Both andreas.both@unister.de
Head of R&D +49 341 65050 24496
Unister GmbH http://www.unister.de
Slide 24 Dr. Andreas Both, Head of R&D, Unister