Post on 15-Jan-2015
description
Business Intelligence
Michael Lamont, ’12
lamont@post.harvard.edu
Decisions
Stra
teg
ic
Tactic
al
Decisions
Decisions
Decisions
Possible to make a good decision
without data technology
Looking at the right data can help you
make better decisions
Decisions aren’t judged on a binary
“good” or “bad” scale.
Decisions
Decision quality is measured on a
gradient scale
Disastrous Excellent
Decisions
Decisions drive companies
Better decisions lead to:
More efficient operation
Higher profitability
Greater customer satisfaction
Companies that make better
decisions are more successful
Business Intelligence
Business Intelligence (BI): Using data
about yesterday and today to make
better decisions about tomorrow
BI makes companies smarter:
The right criteria to judge success
Locating and transforming the right data
Arranging information
Lets management see things more
clearly, and glimpse the future
Limited Resources, Unlimited Decisions
Every organization has to make do with
only what they have, all the time
You can’t hire only the brightest minds
and spend unlimited money on
efficiency
Time is the most precious resource your
company has – it has to move quickly,
not just correctly
Limited Resources, Unlimited Decisions
BI is a powerful ally when a decision is
required
Flexible resource that can be used by
any level of the organization
Limited Resources, Unlimited Decisions
Examples of simultaneous BI tasks:
VP of Sales deciding which markets &
accounts to target to meet sales targets
Product developers deciding which
fragrances to use in future products
Gulf Coast marketing team deciding on
holiday weekend promotions
BI Defined
Lots of very different definitions of
business intelligence exist
BI used to be a marketing buzzword that
got tied to not strictly related
technologies
Every vendor invents a definition that
skews toward their products
Researchers, authors, & consultants all
have their own pet definitions
BI Defined
Business Intelligence is timely, accurate,
high-value, and actionable business
insights, and the work processes and
technologies used to obtain them.
There’s no “magic list” of processes or
software that constitute BI
BI is high tech, and thus always evolving
Every company’s situation is different
Insights
Insights should flow
out of successful BI
projects
An insight is:
A new way to look at
things
A moment of clarity
A path forward
Something that you
didn’t already know
about your company
Business Intelligence
Accurate Valuable Timely Actionable
Accurate Answers
Decisions should be:
Informed by data
Made by subject matter experts
Based on hard information
For BI to be valuable, it has to:
Reflect objective reality
Adhere to strict standards of correctness
Accuracy is a core attribute of BI
insights
Accurate Answers
BI is subject to “Garbage In, Garbage
Out” (GIGO) rule
Business Intelligence Process
Inaccuracy Example
Sales exec sees a region lagging behind
Senior executives adjust sales process
(and personnel) in that region
What if the insight was wrong?
Some sales offices were incorrectly
allocated to neighboring region
Sales volume wasn’t correctly allocated
Actions taken were less than helpful –
may have made things worse
Accurate Answers
Accuracy is also important from a
political perspective
BI can’t have a real impact unless
people trust it
BI insights can be
Surprising
Counterintuitive
Threatening to some groups/managers
Accurate Answers
Any error, no matter how small, is going
to be used to call into doubt every
conclusion pulled from the data
BI has to be as accurate as possible to
protect its reputation from skeptics
Inaccurate BI insights are worse than
useless – they’re damaging
One bad BI experience will keep it from
ever being trusted again
Valuable Insights
Not all insights are created equal
Reporting that people who buy peanut
butter also buy jelly isn’t much of an
insight
BI should produce information that can
have a real impact on a business
Valuable Insights
Impact of valuable BI insights can be:
Reduced costs
Increased sales
Operational efficiency
Other positive factors
High-value insights aren’t usually
deducible
Insights aren’t always obvious, but can
have huge impact
Valuable Insight Example
Walmart analysis of most popular
products after severe hurricane damage
Valuable Insight Example
Timely Information
Die Geist der Treppe – the “Spirit of the
Staircase”
Delivering facts late in a debate keeps
them from mattering
Information delays in business can have
the same level of impact
Timely Information
Information delays come in many forms:
Workflow (not refreshing data frequently
enough)
Technology (lack of computational power
and efficiency)
Unexpected logistics issues
The time taken for each step in a BI
process, added together, has to be short
enough to make results useful
Timely Information
Timeliness is a required part of useful
insights
High quality BI processes need current
information
Analysis products need to be provided
to decision makers in time for them to
consider all courses of action
Actionable Conclusions
Insights are worthless if they can’t be
acted upon
Non-actionable insights:
Major competitors should instantly cease
operating
Factories should be 20 years newer
Decision support tools will happily find
non-actionable insights if you let them
Actionable Conclusions
An insight is actionable if there’s a
reasonable way to take advantage of the
situation
Conclusion Action Result
The Value of BI
Links information with action
What’s the real value returned from an
investment in BI tools and processes?
Promoting and supporting better
decision making habits
The BI Cycle
Raw Operational
Data
Business Insights
Take Action
Measure Results
The BI Cycle
Companies that follow the cycle have a
rational decision making process
Business Intelligence supports the cycle
Obtain insights from operational data
Good insights can be applied to decision
making process
Decisions lead to actions, and improved
operational results
Cycle repeats and decisions are refined
Trends
BI continues to increase in importance
to both large enterprises and SMBs
BI is flexible, and responsive to
technological advances
Trends
BI projects are originating outside IT
departments
IT used to be the only group that knew what
data and analyses were available
Executives and other decision makers have
gotten comfortable with BI
Trends
Delivery of analytics to desktop (and
mobile devices)
Premier vendors can round-trip data to
standard Office applications
Excel includes advanced analytical tools
Trends
Data access is becoming more dynamic
and approaching near-real-time
BI systems of the future will be able to
directly pull data from operational systems
Computational efficiency of BI systems is
constantly increasing
Conclusions
Business Intelligence isn’t just about
computing
Requires a corporate culture that
supports data-driven decision making
Business managers must promote data-
based decision making
IT has to support the tech behind BI at
all levels of the company
Conclusions
BI gives you new tools and perspectives
Lets you ponder what-if questions
Decision makers have to know how to
ask the right questions
No set rules for determining the “right”
reports and analytics for a particular
company
Conclusions
The right people have to be in the right
positions for BI to work
BI is a commitment to rational decision
making processes
Must be supported at all levels of the
company, by both managers and IT
Michael Lamont, ’12
lamont@post.harvard.edu