Friday, November 1, 2013

Unlocking the Big Promise of Big Data


Rotman Management Magazine - Fall 2013


If the questions and challenges around big data can be dealt with effffffectively, the opportunities ahead are extraordinary.

by Anita McGahan

‘BIG DATA’ IS EVERYWHERE. Organizations of all types — large companies,
entrepreneurial startups, governmental agencies and
NGOs — have been encouraged to embrace the detailed, micro level
data generated by digital devices to become more efficient.
Indeed, the potential of big data to improve organizational productivity
Is significant, but that is only the tip of the iceberg.

In this article, I will argue that the most important opportunities
arising from big data lie in addressing problems such as
climate change, the fragility of our financial system, epidemic
disease, widespread corruption, disenfranchisement of the poor,
and the depletion of cheaply-extractable minerals, to name just
a few of the most important problems of our time. In short: the
‘big promise’ of big data lies in making headway against the most
significant management problems of our century.

The Need For a Big Data Strategy
What is big data? The prevailing definition is that a dataset is ‘big’
if it is too large to handle with widely available, conventional software
programs such as Excel, and thus requires specialized analytics.
Big data typically arises from digital devices, such as cellphones,
GPS systems and terminals on the Internet. Examples
include ‘information on every address that a particular taxicab
in NYC traveled to over the past year’; ‘the number of minutes
that individual users spent considering a stock beforemaking an
online purchase’; ‘the bacterial characteristics of beachwater according
to daily tests’; and ‘the identities of the purchasers of bus
tickets from remote rural communities into Mumbai’.

Analysts frequently point out that big data raises as many
questions as it answers: does the available micro-behavioral data
accurately represent the broader decision framework in the situation?
What questions can be accurately addressed by the data?
Who should control datasets and the insights that they generate?
Once a decision is made, how should an organization respond?
There is no shortcut through the challenges of organizational
change based on insights generated from big data. All of these issues
suggest that executives will be well-served by developing a
big-data strategy rather than by reacting opportunistically to the
analyses made possible as it is uncovered.

Strategy is also essential for unlocking solutions enabled by
big data to the most important social problems of our time. Problems
such as climate change, financial instability and epidemic
disease are intractable because they are involved with and are
at least partly remediated by large public systems that cannot be
abandoned — or even put at risk of outage. For example, public
utilities and infrastructure simply cannot support experimentation
that puts their core operation at risk. Thus, the importance
of strategy here is even greater than it is for private corporations
that can tolerate experimentation in the core.

Leaders who seek to unleash the power of big data in the
interests of addressing the most important problems of our time
need a strategy for triaging analytical opportunities, anticipating
the innovations that the analysis will compel, and for overcoming
resistance to change in the face of legitimate claims that current
systems cannot be abandoned without a seamless and responsible
transition to new ways of doing things.

Any strategy for tackling the most important problems of
our time also has to deal with the complexities of implementation
over governmental agencies, private actors and the public.
In short, big data carries the potential to inspire and create new
resources, decision tools, communication vehicles, and coordinating
devices; but realizing their value requires doubling down
on thinking through the implications for a safe, equitable, innovative,
responsible and peaceable transition to new ways of governing
public resources.

Big-Data Tools for Creating Public Value
The first step in an effective big-data strategy involves identifying
analytical tools for creating information that can enable innovative
solutions to public problems. Figure One contains eight
examples of tools that are already in use. Many of these tools are
interrelated, and the list is far from comprehensive. Yet a brief
review of the public processes that have been addressed through
these tools reveals two important sets of issues that must be
addressed for a big-data strategy to be effective: classical challenges
that have long been elements of public processes, such
as accounting for absent voices in public dialogue; and antecedent
challenges that are specifically tied to the use of big-data
tools. Let’s look more closely at each in turn.

CROWDSOURCING. In 2001, a television show called ‘Who Wants to
be a Millionaire’ hosted by Regis Philbin took over the airways
in North America. The show required its contestants to answer
a number of multiple-choice questions successfully. One tool
available to each contestant was to ask the audience for advice
— effectively ‘crowdsourcing’ the answer to the multiple-choice
question at hand (the game allowed a contestant only one instance
of use of this opportunity per game). Despite the specificity
or technicality of the question, the audience almost invariably
produced the right answer.

Crowdsourcing in the public domain typically involves engaging
the public through the Web in a discussion of a potential
problem and its solutions. When crowdsourcing works, it can
cultivate a diversity of potential solutions to a well-defined
problem, such as how to reduce rush-hour traffic on a particular
roadway. A classical challenge in this situation is in accounting
for absent voices. For example, suppose a promising crowdsourced
suggestion for resolving the traffic problem involves diverting
heavy trucks onto an access road. A classical challenge
for a decision maker in this situation involves accounting for
the fact that the majority typically will not consider the property
rights of a minority of people living on the new route who will
be adversely affected by it. But big data also raises another set of
issues around fairly and responsibly considering and responding
to the full range of legitimate alternative routes, as well as other
possible solutions, such as implementing a congestion charge
or restricting truck use to low-density times of day. In other
words, crowdsourcing compels not only a rigorous concern for
old problems, it also creates new ones.

WEB INTERFACES FOR CITIZEN ENGAGEMENT. Over the past 20
years, many small towns around the world have established
websites that invite citizens to offer feedback to local authorities,
and engage in dialogue about local problems. In
some communities, this web-enabled dialogue has deepened
discussion by enabling the development of important
facts that inform questions about the operation of schools,
for example. One classical challenge that such a dialogue creates
is the disenfranchisement of actors that may be critical
in the response to such suggestions, such as an elementary
school’s faculty and staff. Many town managers now regularly
meet at conferences in which they share tips, tactics and
strategies for avoiding such disenfranchisement. Yet the enablement
of web-based tools also creates another problem:
the potential for framing and advancing dialogue on issues
that are not well-defined, widely concerning, or actionable
under localized decision structures. When web feedback loses
focus and becomes diffuse, it can generate as many frustrations
as advances.

OPEN INNOVATION PLATFORMS AND GRAND CHALLENGES. In a single
generation, the Gates Foundation and its partners have
advanced science on the health challenges of the poor through
a program of ‘Grand Challenges.’ These programs establish laddered
goals for solving a previously-intractable problem, and
thus decompose problems to break resource constraints. ‘Open
innovation platforms’ similarly allow scientists to collaborate
more effectively by providing a single place where ‘half-baked’ solutions
and partial advances are accessible across the community.
What kinds of challenges tend to arise in the implementation
of grand challenges and open-innovation programs? A
long-standing issue is in overcoming the private incentives of
scientists to keep their work private — a challenge that the Gates
Foundation has addressed by offering large research grants
to scientists willing to share their work. The grand-challenges
process has also exposed idiosyncratic and antecedent
management problems that have required a carefully developed,
strategic response. Primary among them is the failure even
in the achievement of intermediate goals, such as plagued
scientists committed to the development of an HIV vaccine.
Even out-of-the-box successes may create unanticipated demands
on important actors, such as local physicians, who must
be engaged to deliver novel medicines once they are developed.

ONLINE TOWN HALLS AND OPEN-GOVERNMENT BLOGS. Some elected
and appointed leaders have gone beyond enabling inbound Web
tools and turned to online town halls and blogs as mechanisms
for enhancing transparency in their decision processes. For example,
U.S. President Barack Obama conducted the first ‘live
tweet’ online town hall meeting in May, 2009, in which he invited
questions from American citizens. What is the advantage of such
an approach? Obama explained his priority as “open, transparent
governance.” By inviting dialogue, public leaders may become
more inclusive and transparent in their decision processes.
A classical challenge: fighting shallowness, maintaining focus,
and preventing the town hall from degenerating into a firestorm
of disconnected complaining. By enabling town halls online, the
scale on expectations for transparency may be unprecedented.
To make good on his promise of transparency, Obama had to
maintain coherency by defining specific subjects and decision
priorities as the topics of online discussion.

SOCIAL MEDIA. Three days after the 2013 Boston Marathon bombing,
the Federal Bureau of Investigation, in collaboration with
the City of Boston and the Commonwealth of Massachusetts,
released grainy pictures of two suspects and asked the public
for help in identification. Within seconds, users of social media
such as Facebook and Twitter had identified the figures as
Tamerlan and Dzhokhar Tsarnaev. In subsequent days and
weeks, the FBI used the social-media accounts of the suspects
to construct theories of the motives, mechanisms, and potential
collaborators behind the crimes.

In general, the big data embedded in social media constitute
an extraordinary stockpile of potential information about relationships,
viewpoints and personal activities. For leaders interested
in developing public policies, analyzing these media may
yield understanding of how important stakeholders will react to
alternative approaches. By anticipating these reactions, leaders
can more effectively craft policy in the first place. A classical challenge
involves negotiating among the interests of various stakeholders,
and subsequently managing the expectations of those
whose interests are compromised. An antecedent challenge to
big data itself: responding to the feedback that social media generates,
and staying the course on a singular, integrated and coherent
plan of action.

CALLS TO ACTION. Change.org distributes petitions for social
change. Sign up, click a few boxes to indicate your interests, and
you get daily or weekly e-mails listing petitions to which you
can sign your name. The petitions are designed to lead to specific
action, such as to compel a company to update its childcare
policies, lobby a county animal-rescue shelter not to euthanize
an animal, and to expose corruption in a regional office issuing
driver’s licenses in India. The big-data opportunities generated
by Change.org, which has the e-mail addresses of advocates
willing to put their names on petitions associated with particular
causes, include the potential to galvanize almost instantly advocacy
for a specific action by a public or private authority.

A big-data strategy on a complex issue such as, for example,
reducing corruption in the administration of driver’s licenses,
requires the careful vetting of the petitions posted by organizations
such as Change.org to validate their legitimacy and to assure
that their framing is sensitive to context. A classical issue
in advocacy is avoiding the dilution of public authority and responsibility
that may arise from a targeted intervention (for
example, corruption in the issuance of driver’s licenses may
become more subtle and entrenched in the wake of a single, targeted
intervention.)

The problem of countermanding the classical challenge
is complicated by a new issue antecedent to big data itself: embracing
variation in the emotional commitment of advocates to
particular causes. Change.org makes signing a petition so easy
that many signatories may not feel as strongly as others about the
causes they support. Triaging the level of advocacy for a step is
as important as the email addresses of the advocates.

REGISTRATION AND TRACKING OF RESOURCES. Many organizations —
including governmental agencies, NGOs and companies — sit
on stockpiles of data that could be transformed into valuable
information. Big data is commercially valuable, but assessing
its value is complicated. The first step lies in understanding
the stockpile as well as the organization’s capacity for analysis
and for commercialization of data.

Accurately accounting even for tangible assets is a significant
challenge in many organizations, especially governmental
agencies, and the challenge is compounded when the assets
are intangible: big data may have a short half-life (the value of
information on last year’s fashion trends fades quickly), be unexpectedly
important (information on the efficacy of a vaccine escalates
in value when an epidemic breaks out), and contain ‘needles
in haystacks’ (such as the identities of the Tsarnaev brothers).
Leaders who seek to use big data for change have to find ways
to avoid pouring analytical resources into fading opportunities
while standing ready to act fast to deploy analytics when the
value of big data is evident.

QUANTIFICATION OF MILESTONES AND TRACKING PROGRESS ON PUBLIC
GOALS. The City of Newark, New Jersey, thrives on big data. Charismatic
Mayor Cory Booker is a tireless tweeter — responding
personally to countless citizen queries, claims, and complaints.
Yet the Booker administration has moved well beyond using digital
technology for communication into a strategy of quantifying
milestones and tracking progress on public goals. The City’s website
contains a data-rich report called Newark 2.0 on the City’s
progress on goals such as preventing missed sanitation pickups,
reductions in aggravated assaults, population growth, and newbusiness
formation.

A long-standing challenge in the reporting of figures such
as Newark’s is in their sensitivity and accuracy: for example,
one way to get criminality statistics down in any city is to reduce
arrests rather than the underlying crime; and the resilience and
robustness of new businesses is as important as their formation.
In its reporting, Newark triangulates on measures to address
these challenges. Even more impressive, however, is Booker’s
ability to manage against the temptation that arises from big
data to look at every issue in quantitative terms: by visiting new
businesses, meeting with police officers, and staying close to his
constituents, Booker works to complement the insights that big
data offers with the subtle understanding that only qualitative
interventions can deliver.

In closing
Big data tools carry unprecedented promise, but they also create
a slew of challenges that must be confronted for a strategy to be
effective. The examples discussed herein illustrate some of the
core issues, which include:
·       defending against analytical overload;
·       finding kernels of value;
·       making insights actionable;
·       meeting the expectations of contributors;
·       managing the emotional overload that arises from expression;
·       and protecting privacy.

Big data also raises some fundamental questions of governance:
who owns information extracted from data about your behaviour,
attitudes and relationships? And who decides what can be done
with that information?

If these questions can be answered effectively, the opportunities
ahead are extraordinary, because buried in this data is information
on how to inspire, create, and communicate effectively
to compel massive action against the most important challenges
of our time.

By embracing big data, organizations of all types can re-invigorate
their approaches to value creation by investing to understand
how to satisfy customers efficiently as well as create jobs
that employees love. Public leaders can use big data to understand
and reinforce the vibrancy of local communities; reformers
and advocates can compel conservation, root out corruption,
and advance civil rights; and law-enforcement agents can use big
data to understand disenfranchisement and disaffectation and
thus prevent tragedy. In short, big data makes possible change at
the scale of our most important public problems.
         
 

Anita McGahan is Associate Dean (Research), the Rotman
Chair in Management, Director of the PhD Program and
Professor of Strategic Management at the Rotman School
of Management, with a cross-appointment to the University
of Toronto’s Munk School of Global Affairs. Rotman faculty
research is ranked in the top ten worldwide by the Financial Times.

No comments:

Post a Comment