McKinsey MassTLC Big Data Seminar Keynote - February 28, 2014
1. Sastry Chilukuri
Partner, McKinsey & Company
Strategy and technology for
Pharma, Med Device and Government
BS, Indian Institute of Technology (IIT);
MBA, Kellogg
5. 4
295 exabytes
If you stacked a pile of CD-ROMs
on top of one another until you’d reached
the current global storage capacity
for digital information …
… it would stretch
80,000 km
beyond the moon
10. 9
10x more servers
75x more files
50x more data
By 2020
IT departments
will be looking after
11. 10
… than in all of history prior to thatBig Data and
Advanced Analytics
3,5oo,ooo,ooo,ooo,ooo,ooo,ooo
2010 2012
o Bang
We have produced more data
in the last two years …
12. 11
Lessons from the leaders
Transformation journey1
Open data2
Organization and talent3
Frontline adoption4
13. 12
Achieving transformational impact is a long journey
011998 99 2000 02 08 0906 201105 1003 0704
Net profit
EUR mn
Tesco personal
finance launch
Tesco.com launched
Clubcard relaunched with key fobs
Coupons@Till
Roll out to int’l markets
£1billion given back to customers
Shopper panel
launch
Promotions cut by 1/4
Customer perception improves
100%
data
Lifestyles segmentation
Assortment
tool
“Value index” integrates price and promotion
indices
Macro space
optimisation
All range reviews use substitutability analysis
Promotions data
Finest launched
All promotions post-evaluated
Standard reporting of customer
insight KPIs and analytics
SOURCE: Bloomberg; interviews; annual reports; analyst reports
14. 13
▪ Gender
▪ Age
▪ Household/billing address
▪ Number of people on plan
Multiple stages along the journey – Telco Example
Factual identity data
Level 1
Level 2
Level 4
Level 5
Level 3
▪ Billing preferences –
paper, online, mobile
▪ Payment history and type
(credit, debit, cash)
▪ Credit/history score
Billing data
▪ POS data – location, time,
amount, product
▪ Channel preferences and
behaviour (through e.g.,
Clickfox)
▪ Online channel behaviour
(sales and service)
Usage data
Network data
Internal data sources
▪ Location
▪ Voice, SMS and data
usage (CDR/ DDR))
▪ Data consumption and
type (e.g., online
browsing, app types)
▪ VOD usage (amount,
time, etc) and type (e.g.,
Tbox)
▪ Music consumption (e.g.,
Mog)
Response data
Marketing response data
▪ Contact history
▪ Campaign response
▪ Offered products
▪ Real time IMEI location
▪ Fixed connection location
External data sources
▪ Housing lists
▪ Phone numbers
▪ Further demographic
information
External data
Unstructured data
▪ Social media – e.g.
LinkedIn data allows us to
segment customers by
occupation, recent
promotion, etc…
▪ E-mail
▪ “Big data” insights
▪ E.g., through partnering
with a major retailer
▪ Spending habits –
average spend per
month, preferred payment
types, etc…
Partner data
15. 14
New capabilities and enablers are required to capture
AD&A opportunities
Analytics
service
providers
Public and
internal
data
sources
Data
partner-
ships
Techno-
logy
platform
providers
+
16. 15
Lessons from the leaders
Transformation journey1
Organization and talent3
Frontline adoption4
Open data2
17. 16
Government increasingly providing Open Data
Categories
▪ Product safety data (e.g.,
complaints, recalls)
▪ Census data
▪ Weather data
▪ Geo-spatial data
▪ 40+ countries with open data platforms
▪ Most of data not necessarily prioritized
by value potential
▪ More sensitive information is typically
withheld
Govern-
ment
1
▪ Google search
▪ Social media (e.g., Twitter
feeds, Facebook likes)
▪ Crowdsourcing (e.g., product
features)
▪ Social media and Google search is
mostly openly available data
▪ Companies have started leveraging
crowd sourced feedback
Consumer
behavior/
pre-
ferences
2
▪ Product information offered,
cost, terms, interest, etc.
▪ Statistics on damages/injuries
incurred by customers
▪ Customer reviews of products
▪ Most of this data is proprietary and
considered as competitive asset
▪ Some of this data may become open
in the future
Firms3
Current trendsExample data sources
18. 17
Companies utilizing Open Data for impact
Data sources
Developed predictive models based on
weather forecasts to improve inventory
management
Developed a proprietary methodology to
rate healthcare providers based on publicly
medicare data and data from healthcare
bodies
Developed an algorithm that assess credit
worthiness of customers based on their
social media profiles
Created an online community to crowd-
source product design inputs from
customers
Launched a contest for movie rating
prediction to improve the accuracy of its
existing movie recommendation
Aggregated its transaction data to provide
customer spending trends and patterns to
other companies
Govern-
ment data
Description Impact
$50–100 incremental profit
per account through
optimized line assignment
Increased inventory
velocity and reduced
revenue lost due to stock-
outs
Gathered cost and rating
information on 5,000
hospitals and 16,000
nursing homes
Revenues rose by 400% in
a span of 8 years
Increased recommendation
quality by 10%
Used big data analytic on
proprietary data to provide
relevant intelligence to
customers
Companies
Consumer
behavior
data
Firm data
20. 19
Find the “translators”--people who can bridge different functional areas
Business
Owners
Analytics
IT
Head of
analytics
Data
Analysts
Data
Scientists
Solution
Architects
▪ Ensure best
in class
models
and
algorithms
support
internal
customers
▪ Drive the design and
execution of the
overall data and
analytic strategy
▪ Provide link across IT,
analytics, and business
▪ Solid under-
standing of
statistics and
analytics
which can
leverage into
business
decisions
▪ Ensure
future data
require-
ments and
delivery
roadmap is
robust and
complete
21. 20
Five ways companies are supercharging their talent
Creating partnerships4
1 Making strategic hires
Retaining advanced analytics
talent
5
Sourcing globally2
Training and certification
programs
3
Strategic hires from Yahoo Labs, Google
and Amazon (~20)
Sizable portion of the analytics team located in
India(~40 out of 60)
Set high expectations for all existing analytics
staff (which is vast bulk of analytics talent)
Constantly scan and collaborate with technology
ecosystem players
Dedicates significant senior management
attention to promote/ retain new talent
22. 21
Lessons from the leaders
SOURCE: Source
Transformation journey1
Organization and talent3
Frontline adoption4
Open data2
23. 22
Typical building blocks for accelerating Advanced Analytics impact
Transformed
data model
Modeling
insights
Data modeling
“Black box”
Heuristic insights
“Smart box”
Workflow
integration
Process redesign
Tech-enablement
Adoption
Internal
External
Capability building
Change management
Source
of value1 2 3 4 5
▪ Advanced
statistical analysis
to drive business
insights
▪ Codified heuristics
dispersed in the
organization to
enhance analytics
▪ Management of
large pools of
internal and
external data
▪ External
strategic
partnerships
▪ Developed frontline
and management
capabilities
▪ Proactive change
management and
tracking of adoption
with performance
indicators
▪ Easy-to-use user
interface built in
the platform
▪ Redesigned
processes to
embed rules in the
workflow
▪ Clear articulation
of the business
needs for
advanced
analytics and
assessment of
expected impact
Modeling
insights
Transformed data
model Adoption
Workflow
integrationSource of value
24. 23
Leading organizations adapt their operating model around analytics tools
to fully take advantage of the insights generated
Text
Cutting edge technology
▪ Define the technology
application(s) that will deliver
analytics insights to the
frontline (e.g., call center
management platform)
▪ Identify innovative
alternatives to the current
state to deliver the same
insights in new ways
(e.g., mobile applications)
Structured feedback loops
▪ Explicitly redesign front-line
processes to incorporate
analytics insights into
decision making
▪ Structure future-state
processes to explicitly
facilitate capture decision
outcomes and additional
user feedback (e.g.,
subjective rationale and
supporting facts)
25. 24
Summary
▪ Understand priority use-cases from the business back –
maintain a heatmap
▪ Operate portfolio of opportunities across multiple horizons
▪ Assess and align capabilities to support scale up and
increased levels of sophistication
▪ Drive pilot projects to prove impact, secure funding, and
work out operating model
▪ Focus on expanding talent over time, leveraging external
partners where possible
▪ Focus beyond models to engage frontline users to
achieve sustainable impact
26. Iran Hutchinson, Product Manager and
Software/Systems Architect, InterSystems
Jon Pilkington, Vice President of Products,
Datawatch Corporation @Datawatch
Iran Hutchinson, Product Manager and
Software/Systems Architect, InterSystems
Bob Zurek, Senior Vice
President, Products, Epsilon
Marilyn Matz, Co-Founder and
CEO, Paradigm4 @Paradigm4