1. Four Eras of Analytics
Thomas H. Davenport, Babson College/MIT/Deloitte/International Institute for Analytics
UNC Charlotte Analytics Frontiers Conference
March 30, 2016
1 | 2016 © Thomas H. Davenport. All Rights Reserved

2. Changing the Way Analytics Are Done
Small, structured, static data
Back-office analysts
Slow, painstaking
Internal decisions
Spreadsheets/OLAP/EDW
Descriptive analytics
Big, unstructured, fast-moving data
Rise of data scientists
Data products in online firms
Rise of Hadoop and open source
Visual analytics
“Agile is too slow”
Mix of all data
Internal/external products/ decisions
Analytics a core capability
Move at speed and scale
Predictive and prescriptive analytics
Analytics embedded, invisible, automated
Cognitive technologies
“Robotic process automation” for digital tasks
Augmentation, not automation
1975-?
2001-?
2013-?
2016-?

3. Analytics 1.0│Traditional Analytics
Primarily descriptive analytics and reporting
Internally sourced, relatively small, structured data
“Back office” teams of analysts
Internal decision support focus
Spreadsheets, BI, EDW, etc.
Tom
3 | 2016 © Thomas H. Davenport All Rights Reserved

4. Analytics 2.0│The Big Data Era
Traditional
Analytics
1.0
Primarily descriptive analytics and reporting
Internally sourced, relatively small, structured data
“Back room” teams of analysts
Internal decision support
Spreadsheets, BI
2.0
Big Data
Complex, large, unstructured data
New computational capabilities required
“Data Scientists” emerge
Online firms create “data products”
4 | 2016 © Thomas H. Davenport All Rights Reserved

5. Analytics 3.0│Fast Business Impact for the Data Economy
Traditional
Analytics
1.0
Fast Business Impact for the Data Economy
Primarily descriptive analytics and reporting
Internally sourced, relatively small, structured data
“Back room” teams of analysts
Internal decision support
Spreadsheets, BI
3.0
Seamless blend of traditional analytics and big data
Analytics core to the business
Data and analytics-based products in every business
Industrialized decision-making at scale
2.0
Big Data
Complex, large, unstructured data sources
New analytical and computational capabilities
“Data Scientists” emerge
Online firms create data- based products and services
5 | 2016 © Thomas H. Davenport All Rights Reserved

6. Analytics 4.0│The Cognitive Era
Traditional
Analytics
1.0
Fast Business Impact for the Data Economy
4.0
Cognitive
3.0
Analytics used to make automated decisions
Emergence of “cognitive technologies”
Replacement of human tasks—digital/physical
Augmentation is human focus
Primarily descriptive analytics and reporting
Internally sourced, relatively small, structured data
“Back room” teams of analysts
Internal decision support
Spreadsheets, BI
A seamless blend of traditional analytics and big data
Analytics integral to running the business
Data and analytics-based products in every business
Industrialized decision-making at scale
2.0
Big Data
Complex, large, unstructured data sources
New analytical and computational capabilities
“Data Scientists” emerge
Online firms create data- based products and services
6 | 2016 © Thomas H. Davenport All Rights Reserved

7. (Cumulative) Skills Across the Eras
4.0
1.0
Traditional
Analytics
Fast Business Impact for the Data Economy
Cognitive
3.0
Natural language processing
Neural networks/deep learning
Event stream processing
Work design
Data integration and curation
Storytelling with data
Business acumen
Statistics
Machine learning
Agile methods
Change management
2.0
Big Data
Experimentation
Data restructuring
Open source coding
Product development
Visual analytics
7 | 2016 © Thomas H. Davenport All Rights Reserved

8. Why Move to Cognitive? Too much data Expensive labor
Humans not good decision-makers
Tedious work
Powerful technologies

9. A Smooth Transition from Analytics to Cognitive
Traditional Analytics
Correlation/
Regression
Natural
Language
Processing
Neural
Networks
Logistic
Regression
Deep
Learning
Text
Analytics
Machine
Learning
Cognitive Technologies

10. Smart People Concerned About the Future of Smart People
“I am in the camp that is concerned about super intelligence…I don’t understand why some people are not concerned.” (Bill Gates)
“The development of full artificial intelligence could spell the end of the human race.” (Stephen Hawking)
“Advancing machine intelligence is the most important problem facing the world today.” (Nobel economist Robert Schiller)
“We will soon be looking at hordes of citizens of zero economic value. Figuring out how to deal with the impacts of this development will be the greatest challenge facing free market economies in this century.” (Michael Malone, Bill Davidow)
10 | 2016 © Thomas H. Davenport. All Rights Reserved

11. But Automation Is a Pain
At a minimum, substantial displacement of human workers
In many cases, humans prefer to be served by humans
Automation is often a race to the bottom
Difficult to combine automation with creativity and frequent innovation
11 | 2016 © Thomas H. Davenport. All Rights Reserved

12. Are Knowledge Workers Next to Be Automated?
18th-19th C.
20th C.
21st C.
Mechanical
Systems
Transactional
Computers
Cognitive/
Analytical
Manual
Labor Jobs
Admin/
Service Jobs
Knowledge
Work Jobs

13. My Answer Is…Yes…and No
Many knowledge work job tasks will be automated
Some knowledge workers will lose their jobs, depressing hiring
8 lawyers where there were 10
There will be a lot of jobs (no one knows how many) working alongside smart machines
Immense productivity gains could fund retraining and redeployment of people
But workers can’t afford to be complacent
13 | 2016 © Thomas H. Davenport. All Rights Reserved

14. Ten Knowledge Work Jobs with Automatable Tasks
Teacher/Professor—online content, adaptive learning
Lawyer—e-discovery, predictive coding, etc.
Accountant—automated audits and tax
Radiologist—automated cancer detection
Reporter—automated story-writing
Marketer—programmatic buying, focus groups, personalized s, etc.
Financial advisor—”robo-advisors”
Financial asset manager—index funds, trading
Programmer—automated code generation
Quantitative analysts—machine learning, etc.
14 | 2016 © Thomas H. Davenport. All Rights Reserved

15. Technologies Driving Knowledge Work Automation
Analytics and big data
Machine learning
Neural networks/deep learning
Rule engines
Event stream/complex event processing
“Cognitive computing,” e.g., Watson
Robotic process automation
Custom integrations and combinations of these in a “cognitive cloud”
15 | 2016 © Thomas H. Davenport. All Rights Reserved

16. Just How Smart Are Smart Machines?

17. The Impact on People: Automation or Augmentation?
Augmentation—smart humans helping smart machines, and vice-versa
People do this by aiding automated systems that are better than humans at their particular tasks, or by focusing those tasks at which humans are still better
The classic augmentation example: freestyle chess
Better than humans or automated chess systems acting alone
Humans can choose among multiple computer-recommended moves
Humans know strengths and weaknesses of different programs
We’ve seen this before: textile machinery, spreadsheets
17 | 2016 © Thomas H. Davenport. All Rights Reserved

18. Five Ways of Stepping
Step in—humans master the details of the system, know its strengths and weaknesses, and when it needs to be modified
Step up—humans take a big-picture view of computer-driven tasks and decide whether to automate new domains
Step aside—humans focus on areas they do better than computers, at least for now
Step narrowly—humans focus on knowledge domains that are too narrow to be worth automating
Step forward—humans build the automated systems
18 | 2016 © Thomas H. Davenport. All Rights Reserved

19. The Five Augmentation Steps for Financial Advisors
Step in—advisors become experts in online advice, and assist clients to use it to their best advantage
Step up—advisors identify the domains most in need of automation, or those already automated needing improvement
Step aside—advisors primarily communicate with clients, but don’t make decisions for them—or work outside investments
Step narrow—advisors identify a narrow client segment or investment type
Build the steps—advisors use their expertise to build robo-advisor systems
19 | 2015 © Thomas H. Davenport All Rights Reserved

20. Where Augmentation Is Likely to Prevail
Knowledge bottlenecks
Veterinary diagnosis and treatment
Need for decision quality and consistency
Insurance policy underwriting
Too much data or content for humans to digest/analyze
Digital marketing, oncology
Cognition currently too expensive for broad application
Investment advice, college education
20 | 2016 © Thomas H. Davenport. All Rights Reserved

21. Oncology, For Example Over 400 types of cancer
Hundreds of oncogenes and tumor suppressor genes—probably biome genes too
Oncology information needs to be integrated with EMR data
Over 75 different drugs for breast cancer alone
Treatment options changing very quickly
Diagnosis and treatment too hard for humans, but machines don’t find it easy either—yet
Need for content integration, holistic medicine
21 | 2016 © Thomas H. Davenport. All Rights Reserved

22. Implications for Organizations
Take an augmentation perspective from the beginning
Start a pilot and develop a platform
Pick the right cognitive technology for your problem
Get good at work design for smart humans and smart machines
Give your people the options and the time to transition to them
Put someone in charge of thinking about this

23. Prerequisites for 3.0 and 4.0 Organizations
No silos of people, data, or technology
Leaders who get it and are willing to engage and invest
Relatively few battles between the business and IT
A strategy that focuses on services and processes, not just the best products
A willingness to pursue this for the long haul

24. Some Concerns About 4.0
We don’t fully understand rapid, automated, interconnected decisions
“Flash crash”
Power outages
Snowstorms and air transportation
Need to model dynamic interactions of human and machine behaviors

25. What Are We Waiting For?
The technology is good, and getting better rapidly
Knowledge workers are ready to have their work augmented and to augment smart machines
Cognitive processes need improvement and innovation
We’ll work out the kinks
Start your engines!
25 | 2016 © Thomas H. Davenport. All Rights Reserved