analytics at work - how to make better decision and get better results
TRANSCRIPT
-
8/3/2019 Analytics at Work - how to make better decision and get better results
1/6
2010 Harvard Business School Publishing. Created for Harvard Business Review by BullsEye Resources www.bullseyeresources.compowered by
Aay a W
hoW to mAke better decisionsAnd get better results
featuring t davp
February 2, 2010
key learning summary
in collaboration with
-
8/3/2019 Analytics at Work - how to make better decision and get better results
2/6
HARVARD BUSINESS REVIEW WEBINARS
2010 Harvard Business School Publishing. Created for Harvard Business Review by BullsEye Resources www.bullseyeresources.com. 1www.hbr.org
Analytics at WorkHOW TO MAKE BETTER DECISIONS AND GET BETTER RESULTS
Tom Davenport, Presidents Chair in Information Technology and Management, Babson College
Angelia Herrin, Editor for Research and Special Projects, Harvard Business Review
OVERVIEW
Most organizations have massive amounts of data but fail to
use it in a meaningful way. But with the proper analytical
capabilities, culture, and business processes, organizations
can use analytics to achieve their desired resultmaking
better decisions.
Organizations can benefit from identifying their key strategic
and tactical decisions, assess how well they are doing at
making these decisions, determine which ones can be made
better, and then institutionalize more analytical decision-
making processes.
CONTEXT
Professor Davenport discussed the key concepts from his
latest book titledAnalytics at Work: Smart Decisions,
Better Results.
Dr. Davenports previous book, Competing on Analytics,
focused on companies using analytics to create
competitive advantage. In contrast,Analytics at Work is
designed to help any organization become more analytical
and fact-based. This book also emphasizes the important
linkage between analytics and decision making.
KEY LEARNINGS
There is much wrong with decision making.
In many organizations, there are huge investments in data
warehousing, ERP, and reporting, but this data isnt used to
make better decisions. Bad decision processes and outcomesabound. The body of knowledge on good decision making is
often ignored, and decisions often take too long. There is an
over-reliance on intuition and an under-reliance on data and
analytics. There is also little measurement of decision
processes and outcomes and little accountability for
decisions.
Both decision outcomes and processes areoften bad.
Tom Davenport
From the financial crisis to the decisions to invade Iraq and to
stay in Vietnam, there are numerous examples of bad
decision processes and outcomes in both the private and
public sector. And, even though massive amounts of dataexist, organizations continue to have bad decision processes
and make bad decisions. (While a good decision process
doesnt guarantee a good outcome and bad process doesnt
assure a bad outcome, there is a correlation.)
The good news: there are significant opportunities to improve
decision making. Analytics and algorithms can lead to better
decisions, as can the wisdom of crowds. Use of behavioral
economics can improve decision making and some decision
making can be automated. However, to date most
organizations havent taken advantage of the opportunities
that exist to improve their decision making.
A model exists for making better decisions.
The model shown below is designed to deliver better
decisions. It starts with a foundation of analytical capabilities,
creates an organizational context, delivers better decisions,
and then entails a systematic review to continuously improve
the decision process.
-
8/3/2019 Analytics at Work - how to make better decision and get better results
3/6
Harvard Business ReviewWebinarsAnalytics at Work: How to Make Better Decisions and Get Better Results
2010 Harvard Business School Publishing. Created for Harvard Business Review by BullsEye Resources www.bullseyeresources.com. 2www.hbr.org
Analytical Capability
Professor Davenport has identified organizations at five
stages in developing their analytical capabilities. This ranges
from the analytically impaired (Stage 1) to analytical
competitors (Stage 5).
The capabilities required for any organization to become
more analytical follow the D.E.L.T.A. model.
Data. This is the prerequisite for analytics. At a minimum,
data must be clean, common, integrated, and accessible in a
central data warehouse. Organizations can realize advantage
by having data that competitors dont. This entails
measurements that are new, distinctive/proprietary, and
important. For example, Marriott has proprietary metrics on
revenue optimization and Harrahs measures employee
smile frequency, which predicts customers experiences.
Collect data in areas that others havent
addressed and then apply this data analyticallyin decision making.
Tom Davenport
Enterprise. To become more analytical, organizations must
go beyond managing data locally or in silos. Successful
analytical competitors manage their data and analytics
program at an enterprise level. They create enterprise-wide
analytical capabilities and invest in enterprise-scale analytical
technologies.
Leadership. This is the most critical trait of analytical
companies, and remains extremely rare. Organizations that
become more analytical have leaders who fully embrace
analytics and lead the companys culture toward fact-based
decision making. Quotes of such leaders include, Do we
think or do we know? and In God we trust; all others bring
data.
Targets. With limited analytical resources, analytical
organizations pick a primary strategic target for their initial
analytical efforts (such as marketing or supply chain) as well
as a secondary target. Over time, the use of analytics and
analytical decision making will expand in an organization.
But long-term success starts with a specific strategic
application.
Analysts. An organization cant become more analytical
without analytical people. The types of analytical talent
required include: 1) Champions, who lead analytical
initiatives (perhaps 1% of the organization); 2) Professionals,
who can create new algorithms (5-10%); 3) Semi-
professionals, who can use visual and basic statistical tools(15-20%); and Amateurs, who use spreadsheets (70-80%).
Organizations need each of these types of analysts.
A table on Exhibit 1 fromAnalytics at Work shows how
organizations progress from Stage 1 to Stage 5 of each of the
success factors in the D.E.L.T.A. model.
Organizational Context
The context needed to become more analytical includes
creating an analytical culture and having analytical business
processes:
Analytical culture. An analytical culture is one where use of
facts, evidence, and analysis is the primary way of making
decisions. There is still room for intuition, but intuition
should be based on experience and expertise. In analytical
cultures, it is okay to push back by asking, Wheres your
data? When facts are lacking, organizations with analytical
cultures emphasize testing and learning, and they focus on
action after analysis.
Analytical processes. Becoming more analytical also entails
thinking very analytically about an organizations business
processes. This entails mapping out key processes, such as
the ordering processes, and understanding all possible steps
in the process. Analytics can be used to understand which
customers are most valuable and what actions might be taken
to improve the process. Today it is rare for organizations to
be highly analytical regarding their business processes.
-
8/3/2019 Analytics at Work - how to make better decision and get better results
4/6
Harvard Business ReviewWebinarsAnalytics at Work: How to Make Better Decisions and Get Better Results
2010 Harvard Business School Publishing. Created for Harvard Business Review by BullsEye Resources www.bullseyeresources.com. 3www.hbr.org
Desired Result
The goal for analytics is to make better decisions. However,
while many organizations collect data and some
organizations engage in analytics, few organizations link their
analytics activities to their decision making.
Better decisions are the goal of analytics.
Tom Davenport
In a study of decisions, Professor Davenport found that 90%
of companies made at least some effort to improve a specific
decision. The decisions that organizations looked to improve
tended to be those that were frequent and operational, such
as pricing decisions, targeting decisions, merchandising
decisions, and location decisions.
A system for improving decisions should include the
following steps:
Identify. This entails identifying the organizations key
strategic and tactical decisions, which rarely occurs in most
organizations.
Inventory. Determine how well the organizations key
decisions are being made today. What processes and tools are
being used? Are the key decisions being made intuitively or
analytically?
Intervene. Based on identifying the key decisions and
conducting an inventory of how they are being made,
determine if a decision intervention is requireddoes the
organization need to change how certain critical decisions are
being made? Are better people, processes, and tools required?
(Use of analytics is the most common intervention that
organizations make to improve their decision making).
Institutionalize. Allocate resources (people and technology)
and create processes to institutionalize how the
organizations key decisions are made.
Systematic Review
An important step in improving an organizations analytical
capabilities is to close the loop by engaging in a review of an
organizations key decisions. Many successful organizations
look back at all major decisions to assess the quality of the
decision process and the outcome. They look closely at any
errors that are made and seek to rectify these mistakes in
OTHER IMPORTANT POINTS
Building organizational support. Getting leadership support
for analytics and more analytical decision making can be a
challenge, yet leadership support is critical. Find a senior
executive who is analytical and enlist this person as a partner
in getting the analytical movement going.
The role of intuition. Intuition can play an important role in
generating the hypotheses that analysts investigate and in
deciding when analysis may not be appropriate.
Preventing slow decisions. Organizations can measure how
long decisions take and if they are taking too long, can
investigate to understand why, and can reengineer the
decision process.
Basing KPIs on analytics. In many organizations key
performance indicators (KPIs) are set arbitrarily. By using
analytics, KPIs can be established based on facts.
Analytical measures and incentives. Some organizations are
beginning to measure and reward managers for their decision
processes; not just for the outcomes of their decisions, which
can be delayed by many years.
Analytics for small businesses. Analytics is not just for large
enterprises; small organizations can also use analytics to
make more fact-based decisions. Technology is not the
barrier. Software as a service is increasingly making
analytical tools available to small businesses. The biggest
obstacle is analytical people. However, even such people canoften be rented.
New analytical organization. A new peer-based research
organization has been formedthe Inter-national Institute
for Analytics (IIA). IIA is a community of analytics
practitioners that will conduct research and gather
information on how real-world organizations are deploying
analytics. It is a way to advance the profession of analytics
and analytical decision making. All individual members who
join IIA in February will receive a complimentary
autographed copy of Analytics at Work. Go to
www.iianalytics.com to learn more or join.
Recommended reading. In addition to his own books,
Professor Davenport suggested reading How We Decide by
Jonah Lehrer and Nudge by Richard Thaler and Cass
Sunstein.
-
8/3/2019 Analytics at Work - how to make better decision and get better results
5/6
Harvard Business ReviewWebinarsAnalytics at Work: How to Make Better Decisions and Get Better Results
2010 Harvard Business School Publishing. Created for Harvard Business Review by BullsEye Resources www.bullseyeresources.com. 4www.hbr.org
EXHIBIT 1
-
8/3/2019 Analytics at Work - how to make better decision and get better results
6/6
Harvard Business ReviewWebinarsAnalytics at Work: How to Make Better Decisions and Get Better Results
2010 Harvard Business School Publishing. Created for Harvard Business Review by BullsEye Resources www.bullseyeresources.com. 5www.hbr.org
BIOGRAPHIES
Tom Davenport
Presidents Chair in Information Technology and Management, Babson College
Tom Davenport holds the President's Chair in Information Technology and Management at Babson College. His books and
articles on business process reengineering, knowledge management, attention management, knowledge worker productivity, and
analytical competition helped to establish each of those business ideas. Over many years he's authored or co-authored nine
books for Harvard Business Press, most recentlyCompeting on Analytics: The New Science of Winning (2007). His next book,
Analytics at Work: Smarter Decisions, Better Results, will be published in February 2010. Davenport has authored fifteen
articles forHarvard Business Review. His byline has also appeared for publications such asSloan Management Review,
California Management Review, Financial Times, Information Week, CIO, and many others.
Davenport has an extensive background in research and has led research centers at Ernst & Young, McKinsey & Company, CSC
Index, and the Accenture Institute of Strategic Change. Davenport holds a B.A. in sociology from Trinity University and M.A. and
Ph.D. in sociology from Harvard University.
Angelia Herrin
Executive Director of Project Development, Harvard Business Review
Angelia Herrin is executive director of business development at Harvard Business Review. At Harvard Business Review, Herrin
oversaw the re-launch of the management newsletter line and established the conference and virtual seminar division for
Harvard Business Review. More recently, she created a new series to deliver customized programs and products to organizations
and associations.
Prior to coming to Harvard Business Review, Herrin was the vice president for content at womenConnect.com, a Web site
focused on women business owners and executives.
Herrins journalism experience spans twenty years, primarily with Knight-Ridder newspapers and USA Today. At Knight-
Ridder, she covered Congress, as well as the 1988 presidential elections. At USA Today, she worked as Washington editor,heading the 1996 election coverage. She won the John S. Knight Fellowship in Professional Journalism at Stanford University in
198990.
The information contained in this summary reflects BullsEye Resources, Inc.s subjective condensed summarization of the applicable conference session. There may bematerial errors, omissions, or inaccuracies in the reporting of the substance of the session. In no way does BullsEye Resources or Harvard Business Review assume anyresponsibility for any information provided or any decisions made based upon the information provided in this document.