5. • Big Data: The Management
Revolution
• Data Scientist: The Sexiest Job
of the 21st Century
• Making Advanced Analytics Work
for You
Hot Off the Press
Table of Contents: October 2012
RAY GENSINGER 2012 5Harvard Business Review, October 2012
6. The New Economy?
RAY GENSINGER 2012 6
“Data is the new oil; it is both
valuable and plentiful but useless
if unrefined” – Clive Humbly, visiting professor of
integrated marketing, Northwestern University
7. Industry Analytics: Baseball
RAY GENSINGER 2012 7
• Commentator:
“ It’s a cold one tonight…Joe
Mauer is up and is facing Phil
Hughes. Joe has yet to get on
base against Hughes so he’s
about due…That could be unlikely
though given the temperature.
Joe’s OBP is about 0.125 lower
when the temperature is less than
50 degrees out.”
Sunday Night Baseball
• Infrastructure:
• Every play of every game
captured
• Sabremetrics
• Detailed meteorological data
• Situational metadata
• Runs in the season
• Men on base
• Current count
8. Industry Analytics: Baseball
RAY GENSINGER 2012 8
Historically
• Gut reactions of scouts
• Performance to date
• Batting average
• Hits only
• Compensation based on
individual performance at the
plate
Oakland A’s: Moneyball
Innovation
• Winners score more runs
• Runs score after you have a
base runner
• Who gets on base the most
• On base percentage
• Hits
• Walks
• Analysis of the interactions and
performance of players in relation
to each other
9. Industry Analytics: Baseball
RAY GENSINGER 2012 9
• Season ticket prices set for entire season
• Individual game tickets priced prior to season starting based on last seasons
data
• Popularity of opponent
• Attendance of last seasons games
• Once season starts the game prices vary daily
• Popularity of home and away teams
• Streaking teams or players
• Weather trends
Minnesota Twins: Variable Ticket Pricing
10. • Data rich
• Information intensive
• Asset intensive
• Time dependence
• Quality control essential
• Dependence on distributed
decision making
Healthcare and Analytics
RAY GENSINGER 2012 11
http://www.madisonshope.com/images/Latest/Madison%20in%20ICU%20with%20all%20her%20support%20equipment.jpg
11. • Descriptive analytics provides
simple summaries about the
sample and about the observations
that have been made. Such
summaries may be either
quantitative, i.e. summary
statistics, or visual, i.e. simple-to-
understand graphs. These
summaries may either form the
basis of the initial description of
the data as part of a more
extensive statistical analysis, or
they may be sufficient in and of
themselves for a particular
investigation.
Descriptive Analytics
12http://quiqle.info/8731-how-to-read-a-bell-curve.html
12. • O/E Readmissions .99
(0.9 threshold)
• Optimal asthma care 56.7%
(42.3% threshold)
• Optimal diabetes care 37.4%
(32.9% threshold)
• Breast cancer screening 74.5%
(75% threshold)
• Patient rating of care 72.8%
(71.3% threshold)
• Kevin Love scores 26 points per game
• Kevin Love gets 13 rebounds per game
• Kevin Love plays 39/48 min per game
• Kevin Love has a .448 shooting
percentage
• Ricky Rubio hands out 8 assists per
game
• Ricky Rubio plays 34/48 min per game
• Timberwolves final record was 26-39
(.400)
Descriptive Analytics: Example
Sports Analogy (published) Healthcare Analogy
13
13. • Predictive analytics is an area of
statistical analysis that deals with
extracting information from data
and using it to predict future trends
and behavior patterns. The core of
predictive analytics relies on
capturing relationships between
explanatory variables and the
predicted variables from past
occurrences, and exploiting it to
predict future outcomes. It is
important to note, however, that
the accuracy and usability of
results will depend greatly on the
level of data analysis and the
quality of assumptions.
Predictive Analytics
14http://www.simafore.com/blog/bid/78815/Does-LinkedIn-group-growth-mirror-
Predictive-Analytics-hype-cycle
14. • 50% of all readmission that are
diagnosed with heart failure and
go home on more than 10
medications, but lacking an ACE
or ARB
• Pre-diabetic patients from NE
Minneapolis age 25-35 are 77%
more likely to go on to diabetes
than typical pre-diabetics
• Kevin Love increases his
shooting percentage by 0.15
when Rubio is on the floor
• Winning percentage increase
from 0.4 to 0.45 when both Love
and Rubio play a minimum of 35
minutes
• Other players shooting
percentage drops 0.08 when
Love and Rubio are on the floor
together
Predictive Analytics: Example
Sports Analogy Healthcare Analogy
15
15. Prescriptive Analytics
16
• The final analytic phase is Prescriptive Analytics.[3] Prescriptive Analytics
goes beyond predicting future outcomes by also suggesting actions to
benefit from the predictions and showing the decision maker the implications
of each decision option.[4] Prescriptive Analytics not only anticipates what will
happen and when it will happen, but also why it will happen. Further,
Prescriptive Analytics can suggest decision options on how to take
advantage of a future opportunity or mitigate a future risk and illustrate the
implication of each decision option. In practice, Prescriptive Analytics can
continually and automatically process new data to improve prediction
accuracy and provide better decision options.
16. • Pre-diabetic patient from NE
Minneapolis
• Age 25-35
• Engage in a wt. loss program and
dietary modification
• Initiate ACEi therapy if hypertensive
• Arrange group counseling and therapy
• Age >35
• Consultation endocrinology
• Assign care coordinator
• Purchasing Pattern: March
• Cocoa Butter
• Larger purse
• Vitamin supplements including
zinc and magnesium
• Blue rug
• Marketing plan
• Customized direct mail adds for
baby products
• 87% likelihood of delivering a
baby boy in August!
Prescriptive Analytics: Example
Marketing Analogy Healthcare Analogy
17
18. Questions To Be Addressed
RAY GENSINGER 2012 20Adapted from Analytics at Work. Davenport, Harris, and Morison
What happened?
(reporting)
What IS happening
now?
(alerts)
What WILL happen?
(extrapolation)
How and why DID it
happen?
(modeling, experiment)
What’s the NEXT best
action?
(recommendation)
What’s the best/worst
that CAN happen?
(predict, simulate)
19. When Are Analytics NOT Helpful?
RAY GENSINGER 2012 21
• When there is no time………………………………..Answers needed now
• When there is no precedent…………………………Falling housing prices
• When history is misleading
• Highly variable history…………………………………Major economic turmoil
• Highly experience decision maker (wisdom)………The proven expert
• Immeasurable variables………………………………Emotion, family situation
20. • Carelessness (Mars Orbiter)
• Failing to consider analysis and
insights
• Failing to consider alternatives
• Waiting too long to gather data
• Postponing decisions
• Asking the wrong questions
• Starting with incorrect
assumptions (housing prices will
continue to rise)
• Finding an analytic that identifies
the answer you were seeking
• Failing to fully understand
alternative of data interpretation
Decision Making Errors
Logic Errors Process Errors
RAY GENSINGER 2012 22
21. Organizational Analytic Readiness
RAY GENSINGER 2012 23
• Analytically Impaired
• Lacking skills, data, or leadership
• Localize Analytics
• Disparate glimmers lacking in coordination
• Analytic Aspirations
• Willingness but lacking in a DELTA element
• Analytic Companies
• Tools and people but hasn’t turned content into a competitive advantage
• Analytic Competitors
• Uses knowledge gained to compete and succeed
Five Stages of Development
23. Analytics Success
RAY GENSINGER 2012 25
• The Trouble with cubes
• Unstructured data is a lot like panning for gold, first you sift a lot of dirt
• Uniqueness
• What is it that you have that NOBODY else has
• Nike+ running sensors
• Best Buy Reward Zone
• Health Insurance Companies
• Integration through key identifiers
• Quality is less necessary secondary to the volume of data available
Data
24. Understanding the “Mass” of DATA
• Volume
• World generates 2.5 exabytes of internet traffic each day (zetabyte annually)
• One second of traffic today equals the totality of traffic in all of 1992
• Exabyte
• 1000 petabytes
• 1000 terabytes
• 1000 gigabytes
• 1000 megabytes
• 1,000,000,000,000,000,000 bytes = 1 quintillion bytes
26RAY GENSINGER 2012
28. Analytics Success
RAY GENSINGER 2012 30
• Integration of data from across the organizational silos
• Disparate data isn’t local or independent, it is FRACTURED
• Duplication of resources, services, licenses, subscriptions
• IT Leadership
• Guide the work that matters
• Create an infrastructure that can be widely leveraged
• Share a roadmap with both short and long term success strategies
Enterprise
29. Analytics Success
RAY GENSINGER 2012 31
• There has to be a recognizable name and title behind the strategy; preferably
a CxO
• Hire smart people and recognize them for what the contribute
• Demand data and analysis for all decisions to be made
• Balance analysis, experience, wisdom
• Invest in the necessary infrastructure as a strategic imperative along with
any other high profile strategy
Leadership
30. Analytics Success
RAY GENSINGER 2012 32
• What is it PRECISELY that you would like to achieve?
• Retail:
• Inventory management, price optimization
• Hospitality
• Customer loyalty
• Healthcare
• Maximize accuracy of initial diagnoses
• Highest value care path…Expected outcomes or better at the lowest cost
• Discover opportunities for differentiation
• Goals
• Eliminate the exodus of patients
Targets: What to Achieve
31. Analytics Success
RAY GENSINGER 2012 33
• Complex work streams with many variables or steps
• Simple decisions require absolute consistency
• When an entire service line is in need of attention
• Processes that require complex inputs, connections, and correlations
• Anywhere forecasting is necessary
• Current areas of below average performance
Targets: Where to Achieve
32. Answer the Questions: Set the Targets
RAY GENSINGER 2012 34
What happened?
(reporting)
What IS happening
now?
(alerts)
What WILL happen?
(extrapolation)
How and why DID it
happen?
(modeling, experiment)
What’s the NEXT best
action?
(recommendation)
What’s the best/worst
that CAN happen?
(predict, simulate)
Dr. Surgeon’s cases
start 30-45 minutes
late
Room C, Dr.
Surgeon’s, 30 minutes
behind schedule
Nurses in Dr.
Surgeon’s rooms will
require OT, annual
costs determined
Dr. Surgeon clocks into
the the parking ramp
15 minutes late daily
Schedule a quick case
early morning ahead of
Dr. Surgeon
Start times consistent,
OT drops,
Revenue/case
increases, Dr. Surgeon
quits; either way
finances better
33. Target = Growth
Strategy = Employee Retention
RAY GENSINGER 2012 35Adapted from the “Putting the Service Profit Chain to Work,” HBR, Mar-Apr,
1994.
Internal Services Enhancing Quality for all Employees
Enhance
environment
Flexible staffing
Financial
benefits
Growth
opportunities
Enhanced Employee Satisfaction
Improved
attendance
Employee
retention
Employee
productivity
Enhanced Patient Experience
Consistent
services
Improved
interactions
Empathetic
environment
Service oriented
Improved Patient Satisfaction
Retention
Word of mouth
Social media
accolades
Employer
expectations
Patient Loyalty
Better outcomes
Increased
market share
Revenue growth
Profit
34. Analytics Success
RAY GENSINGER 2012 36
• Analytic Champions: <1%
• Executive decision makers hooked on analysis
• Willing to change the business based on the results
• Analytic Professionals: 5-10%
• PhDs in economics, statistics, research methods, mathematics (or evaluation
studies)
• Programmers and statistical model developers
• Analytic Semi-professionals 15-20%
• MBAs or process improvement experts
• Apply and work the models and theories of the champions and pros
Analysts: Skills and Backgrounds
35. Analytics Success
RAY GENSINGER 2012 37
• Analytic Amateurs: 70-80%
• Knowledgeable consumers of data
• Business managers operating a business unit or help desk staff trying to
anticipate the source of a system error
• The Farm Team
• Curious innovators from around the company
• Ask questions challenging the status quo
• Experiences unrelated to healthcare but knowledgeable about math and
statistics
Analysts: Skills and Backgrounds, cont.
37. • Good to great salary
• Opportunity to create something
new
• Recognition
• Lots of unstructured data
• Autonomy
• Access to “the Bridge”
• Know where to look
How to Catch a Quant
The Really Big Fish You Need the Right Kind of Lure!
RAY GENSINGER 2012 39http://www.dnr.state.oh.us/Default.aspx?tabid=19220
38. RAY GENSINGER 2012 40
Let them know you are interested in how they contribute to the field
Ask them how their work can apply to business challenges
Offer them a challenge to evaluate as part of the interview
Assess their coding/programming skills
Host a competition
Check with your local venture capitalist
Scan through LinkedIn (who do you think created this anyway)
Scan through the “R user groups” (http://blog.revolutionanalytics.com/local-r-
groups.html)
Large universities as well as the unknowns (UT Austin, UC Santa Cruz)
Hang out at Hadoop World (http://www.hadoopworld.com/)
Top Ten Ways to Find a Quant
39. RAY GENSINGER 2012 41
http://www.talend.com/blog/2010/10/14/a-great-hadoop-world-congratulations-to-cloudera/
40. Analytics Success
RAY GENSINGER 2012 42
• Data is a statistician’s crack,
once you have a sample you
can possibly get enough”.
Analysts: Organizing and Developing
Corporate
Division
Analytics
Group
Analytics
Project
Function
Analytics
Group
Analytics
Project
Center of
Excellence
44. References
• http://youtu.be/bVY7OmYqBSY
• Davenport, T., Harris, J., Morsion, R. Analytics at Work: Smarter Decisions, Better Results. Harvard
Business Press. Boston, Massachusetts. 2010.
• http://www.simafore.com/blog/bid/78815/Does-LinkedIn-group-growth-mirror-Predictive-Analytics-hype-cycle
• Adapted from the “Putting the Service Profit Chain to Work,” HBR, Mar-Apr, 1994.
• http://media-cache-ec6.pinterest.com/upload/272327108688186258_0HugpL3c.jpg
• http://quiqle.info/8731-how-to-read-a-bell-curve.html
• http://www.madisonshope.com/images/Latest/Madison%20in%20ICU%20with%20all%20her%20support%20
equipment.jpg
• McAfee, A., Brynjolfsson, E. Big Data: The Management of Revolution. Harvard Business Review. 2012;
90(10):60-69.
• Davenport, T, Patil, DJ. Data Scientist: The Sexiest Job of the 21st Century. Harvard Business Review. 2012;
90(10):70-77.
• Barton, D., Court, D. Making Advanced Analytics Work for You. Harvard Business Review. 2012; 90(10):78-
83.
• http://www.dnr.state.oh.us/Default.aspx?tabid=19220
47RAY GENSINGER 2012