This document discusses data science and business models. It notes that understanding the business problem and delivering value is key for data scientists. Different types of cloud services and business models using machine learning are described, including everything-as-a-service, infrastructure as a service, and software as a service. Standardization is important but data science problems often depend on unique business contexts. Data sources and platforms from AWS, Microsoft, Google, and Dell are also mentioned.
2. classic false move in an immature data culture is “working on the problem where
they have convenient data, without really thinking about the problem”
lessons from my experience
the link to the business and delivering value continuously is
the biggest challenge for data scientists/companies
Business challenge
30.09.17 Frank Kienle p. 2
4. Common Themes Among Successful Data-Driven Startups,
Max Levchin (https://www.youtube.com/watch?v=ylPY7EGrsEE)
30.09.17 Frank Kienle p. 4
data
brokers
e.g. visualize it,
rank it
Share things,
Uber à cars
… à charwomen
--- à daily life
equipment
Lower costs for personal services by data,
Finance, insurance, contracts,
Construction,
Predict it,
operate towards
the future
Model
uncertain
upside
5. Impact on existing business models
30.09.17 Frank Kienle p. 5
Everything-as-a-Service
7. On-Premises
Different types of cloud services
30.09.17 Frank Kienle p. 7
Applications
Data
Runtime
Middleware
O/S
Virtualization
Servers
Storage
Networking
Youmanage
Infrastructure
As a Service
Platform
As a Service
Software
As a Service
8. On-Premises
Different types of cloud services
30.09.17 Frank Kienle p. 8
Applications
Data
Runtime
Middleware
O/S
Virtualization
Servers
Storage
Networking
Youmanage
Infrastructure
As a Service
Applications
Data
Runtime
Middleware
O/S
Virtualization
Servers
Storage
Networking
Youmanage
OtherManage
Platform
As a Service
Applications
Data
Runtime
Middleware
O/S
Virtualization
Servers
Storage
Networking
Youmanage
OtherManage
Software
As a Service
Applications
Data
Runtime
Middleware
O/S
Virtualization
Servers
Storage
Networking
OtherManage
9. Customization, higher costs, slower time to value
Customization vs. Standardization
30.09.17 Frank Kienle p. 9
Standardization, lower costs, faster time to value
10. Value as a Service
30.09.17 Frank Kienle p. 10
§ shift from product-based to software-as-a
service based business models using cloud
computing as the delivery medium.
§ Sooner or later most of the business models
will be subscription based, then the main
focus will be on the value of the service to
your stakeholders.
§ Over time, the move to SaaS has a
commoditization element to it, and the
ability to measure customer value and
desired business outcomes will be true
differentiation. (source: Value-as-a-Service
@Rob Bernshteyn)
11. Challenge for Data Science/AI in value as a service
30.09.17 Frank Kienle p. 11
Standardization, lower costs, faster time to value
§ The shift to software or value as a service
requires standardization
§ Standardization requires a repetitive
problem to be solved
§ Data science problems are often linked to
business specific dependencies
§ A business advantage is defined by a unique
value proposition
§ Every data science/AI service which can be
commoditized will be sooner or later
commoditized and offered as a service
§ Individualized business services will be
build on top of platform as a service or
supportive software as a service offerings
12. Many many companies for different sectors: economy, stocks, weather, global
calendar/event, ….
Example: social media
www.gnip.com
Example: Oracle (https://www.oracle.com/marketingcloud/partners.html)
Data as a Service Provider
30.09.17 12Frank Kienle
13. Overview of data sources
• http://www.knuggets.com/datasets/index.html
Machine learning data
• UCI Machine Learning Repository: archive.ics.uci.edu
Data Shop: the world’s largest repository of learning interaction data
• https://pslcdatashop.web.cmu.edu
For data science: getting Data is not the problem
- Very large flavor of Data Sources
30.09.17 Frank Kienle 13
However, many data are already cleaned for a special focus
14. World wide service platforms: AWS
30.09.17 Frank Kienle 14
AWSoffersfullstackincluding
applicationcentricservices
Example customers
16. World wide service platforms: Google Cloud Platform
30.09.17 Frank Kienle 16
17. The Dell Imperium (On-Premises and cloud services)
30.09.17 Frank Kienle p. 17
DELoffersfullstackincluding
applicationconsultingPivotal
Making Sense of Dell – EMC - VMware https://
a16z.com/2015/10/26/dell-emc-vmware/
18. Business models (SaaS) on machine learning
30.09.17 18
§ www.kaggle.com
platform for predictive
modeling competitions
Focus on learn, work, play
§ A great ressource for
Frank Kienle
19. http://www.skytree.net:
Machine Learning Companies
(attention strongly personal/external opinion)
30.09.17 Frank Kienle p. 19
The claim to have generalized machine learning
models for different use cases is questionable,
the link to business understanding not given in
the examples
Please remember:
80% of your time will be spent in
understanding/cleaning the data and the link to
a business case/business embedding
20. New services to disrupt existing business
https://fleximize.com/paypal-mafia/
30.09.17 Frank Kienle p. 20
21. New Business models on existing platforms
30.09.17 21Frank Kienle
www.uber.com Platform cars
Technology
View:
https://eng.uber.com/tech-stack-part-two/
23. Every products get digitized: àsoftware is eating the world
Examples:
• Fastest growing automotive company: Tesla (run by software engineers)
• Today’s fastest growing telecom company is Skype
• LinkedIn is today’s fastest growing recruiting company
• Amazon Buys Whole Foods (software company buys a retailer)
• General Electric: ‘Bytes will eat machines’ (Forum with Marc Andreessen)
Moores law is way more than just doubling transistor density:
every single day it becomes easier for someone else to
compete with your product
Software is eating the world!
https://a16z.com/2016/08/20/why-software-is-eating-the-world/
2330.09.17 Frank Kienle
24. Impact on existing business models!
it is all about the digital transformation
…‘Digitalization is the use of digital technologies to
change a business model and provide new revenue
and value-producing opportunities; it is the process
of moving to a digital business...
30.09.17 Frank Kienle p. 24
25. Big data to transform business models
30.09.17 Frank Kienle p. 25
Source: Big Data and the Creative Destruction of Today's Business Models (http://www.atkearney.de/)