Closing the Gap Between Strategy and Execution

Abstract

Many markets are affected by the complex interactions of multiple variables: geopolitics, technical innovation, capital market swings, competitive dynamics, shifting consumer preferences and so on. These volatile markets throw out a steady stream of opportunities and threats, and managers can neither predict nor control the form, magnitude or timing of future events with accuracy. In such environments, the traditional linear view of strategy ¿ plan then execute ¿ is woefully inadequate because it hinders people from incorporating new information into action. But instead of thinking of strategy as a linear process, why not consider it as inherently iterative ¿ a loop instead of a line? According to this view, every strategy is a work in progress that is subject to revision in light of ongoing interactions between the organization and its shifting environment. To accommodate those interactions, the strategy loop consists of four major steps: making sense of a situation, making choices on what to do (and what not to do), making those things happen and making revisions based on new information. Reconceptualizing strategy as an iterative loop is simple enough, but putting that new mindset into practice is not. Here, the crucial thing to remember is that discussions ¿ formal and informal, short and long, one-on-one and in groups ¿ are the key mechanism for coordinating activity inside a company. Thus, to put the strategy loop into practice, managers at every level in the organization must be proficient at leading discussions that reflect the four major steps (making sense, making choices, making things happen and making revisions). Companies such as Diageo Ireland, All America Latina Logistica and Onset Venture Services demonstrate that each of the four types of discussions has a different objective that requires a specific tone, supporting information, leadership traits and accompanying tactics.

Full text

Closing the Gap Between
Strategy and Execution
SUMMER 2007 VOL.48 NO.4
REPRINT NUMBER 48412
Donald N. Sull
Please note that gray areas reflect artwork that has been
intentionally removed. The substantive content of the
article appears as originally published.

30 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007
Strategy
Closing the Gap Between
 StrategyandExecution
n an ideal world, managers could formulate a long-term strategy, methodically
implement it and then sustain the resulting competitive advantage. Reality,
however, is rarely so neat and tidy. Technologies evolve, regulations shift, cus-
tomers make surprising choices, macroeconomic variables fluctuate and
competitors thwart the best-laid plans. Thus, to execute strategy as circumstances change,
managers must capture new information, make midcourse corrections and get the timing
right because being too early can often be just as costly as being too late. But how can man-
agers implement a strategy while maintaining the flexibility to roll with the punches?
The first step is to abandon the long-held view of strategy as a linear process, in which
managers sequentially draft a detailed road map to a clear destination and thereafter imple-
ment the plan. This linear approach suffers from a fatal flaw: It hinders people from
incorporating new information into action. How so? First, the linear approach splits the
formulation of strategy from its execution. (Indeed, many business schools still teach for-
mulation and implementation as separate courses.) Thus planners craft their strategy at the
beginning of the process, precisely when they know the least about how events will unfold.
Executing the strategy, moreover, generates new information — including the responses of
competitors, regulators and customers — that then becomes difficult to incorporate into
the prefabricated plan. Second, a linear view of strategy pushes leaders to escalate commit-
ment to a failing course of action, even as evidence mounts that the original strategy was
based on flawed assumptions.1 Leaders commit to a plan, staking their credibility on being
right. When things go awry (the U.S. involvement in Vietnam is a classic example), they find
it difficult to revise their strategy and instead attribute problems to “unexpected setbacks,”
which is just another way of saying new information. Third, a linear approach ignores the
importance of timing. When companies view strategy as a linear process, they sprint to beat
rivals. But rushing to execute a flawed plan only ensures that a company will get to the
wrong place faster than anyone else. Instead, managers need to notice and capture new in-
formation that might influence what to do and when to do it, including the possibility of
delaying as well as accelerating specific actions.
Many managers, of course, recognize these limitations and attempt to work around them.
One approach is to identify big bets up front and then think exhaustively in the planning
process to envision possible outcomes ex ante.2 But managers can rarely identify all the fac-
tors that will end up mattering in the future, let alone predict how events will unfold.
Another approach is to accept the presence of uncertainty, make a best guess on a strategy
based on the data at hand, commit to the strategy and then hope for the best.3 But even
though executives might try to mitigate risk by, for example, diversifying their lines of busi-
ness, the fundamental logic remains: Place your bets and take your chances.
Donald N. Sull is an associate professor of management practice at the London Business School.
Comment on this article or contact the author through smrfeedback@mit.edu.
I
In fast-paced industries,
companies should think
of strategy as an iterative
loop with four steps:
making sense of a
situation, making choices,
making things happen
and making revisions.
Donald N. Sull
SLOANREVIEW.MIT.EdU SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 31
Strategy
There is an alternative. Instead of thinking of strategy as a lin-
ear process, why not consider it as inherently iterative — a loop
instead of a line? According to this view, every strategy is a work
in progress that is subject to revision in light of ongoing interac-
tions between the organization and its shifting environment. To
accommodate those interactions, the strategy loop consists of four
major steps: making sense of a situation, making choices on what
to do (and what not to do), making those things happen and mak-
ing revisions based on new information. (See “About the Research,”
p. 32 and “The Strategy Loop,” p. 33.) These steps can be embed-
ded within formal processes, such as strategic planning, budgeting,
resource allocation or performance management, but they should
also be contained within the myriad informal conversations that
fill out the typical manager’s day. And these discussions should
not be concentrated at the top; they must take place at every level
of the organization. Strategy will remain stranded in the executive
suites unless teams throughout the organization can effectively
translate broad corporate objectives into concrete action by mak-
ing sense of their local circumstances, making
choices on how best to proceed, making things
happen on the ground and making revisions in
light of recent events.
The fundamental advantage of strategy loops
is their ability to incorporate new information
and translate it into effective action. They inte-
grate formulation and execution into a strategic
yin and yang that cannot be separated. They also
explicitly call for ongoing revision as new infor-
mation emerges, mitigating the tendency to
escalate commitment to a failed course of action.
Finally, by breaking time into discrete chunks
(defined by each iteration) and by building in an
explicit step for revision, they increase the odds
that managers will spot changes in context that
open a window of opportunity and will act before
the window closes.
Reconceptualizing strategy as an iterative loop
is simple enough, but putting that new mind-set
into practice is extremely difficult. Here, the cru-
cial thing to remember is that discussions — formal
and informal, short and long, one-on-one and in
groups — are the key mechanism for coordinating
activity inside a company, especially within large
corporations. Thus, to put the strategy loop into
practice, managers at every level in the organiza-
tion must be proficient at leading discussions that
reflect the four major steps (making sense, making
choices, making things happen and making revi-
sions).
All too often, though, conversations at compa-
nies bog down in an endless series of unproductive meetings in
which the usual suspects cover the same ground without making
progress. Frustration mounts as participants “spin their wheels” or
“talk in circles.” To avoid that, managers should start by asking a
simple question: Are we having the right type of conversation?
Specifically, are we trying to make sense, prioritize, make things
happen or revise assumptions? (See “What Are We Talking About?”
p. 35.) Moreover, managers who understand the intricacies of the
four different types of discussions will be better able to translate
understanding into action — and to revise both understanding
and action in light of new information.
Although each type of discussion is simple in principle, they are
all prone to breakdowns in practice. Indeed, the path through the
strategy loop is strewn with pitfalls, but the crucial thing is that
each of the four types of discussions has a different objective, re-
quiring a specific tone, supporting information, leadership traits
and accompanying tactics. (See “Discussions Through the Strategy
Loop,” p. 36 for a high-level summary of those differences.)
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU

Strategy
32 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007 SLOANREVIEW.MIT.EdU
Making Sense of a Situation
The first step of the strategy loop consists of gathering raw data
from different sources to identify patterns from a welter of infor-
mation that is complex, incomplete, conflicting, ambiguous and
of uncertain reliability. The objective is to develop a shared men-
tal model that helps people anticipate how events might unfold.
But the goal of the process should not be accurate long-term
predictions. Instead, people should strive for just enough clarity
to proceed through one iteration of the strategy loop.
To make sense of a situation, managers should establish a tone
of open inquiry rather than advocacy. Teams are most likely to
make sense of novel situations if they dig into the data with an
open mind. In this step, the advocacy of a preconceived interpre-
tation can be dangerous. Consider the Cuban missile crisis.4
While President John F. Kennedy was trying to assess the situa-
tion, his military advisers reflexively advocated invading Cuba, a
course of action they had favored for some time, even though the
specific situation at hand suggested that a military strike could
easily escalate into nuclear war.
Research on effective decision making has found that groups
in rapidly changing markets do best to avoid anchoring too
quickly on a single view.5 In novel situations, the best interpreta-
tion is rarely obvious, and the obvious one is often wrong.
Therefore, the discussion leader must ensure that participants
feel safe to put forth alternative interpretations.6 Kennedy’s team
might have settled on the “obvious” interpretation that Nikita
Khrushchev’s intentions were hostile, but Llewellyn “Tommy”
Thompson, a former ambassador to the Soviet Union, argued
that the Soviet leader probably felt backed into a corner and
might accept a face-saving way to de-escalate the tensions — an
interpretation that proved accurate. (This example also illustrates
the benefit of empathy in making sense of an ambiguous situa-
tion. Thompson knew Khrushchev personally, which helped him
to see the situation from Khrushchev’s perspective rather than
viewing the Soviets as an abstract enemy.)
Instead of passively waiting for divergent views to emerge,
leaders can actively stimulate them. President Kennedy required
his advisers to generate different alternatives to a military strike,
which made it safe for them to discuss the apparently “soft” op-
tions of blockade and diplomatic negotiation — alternatives that
ultimately prevailed, allowing the United States to avoid a nuclear
war. A quick test of whether a team feels comfortable proposing
alternative interpretations is to track the number of framings that
were proposed and seriously discussed.
Conversations to make sense of a situation can, of course,
derail in many ways. The team might cower before a powerful
Over the past decade, I have investigated
dozens of companies in volatile markets.
The core of this research consisted of
comparative case work analyzing how
similar companies in highly uncertain en-
vironments responded to unexpected
opportunities and threats.i The first study
contrasted 10 pairs of established Brazil-
ian firms, in which the focal company
succeeded during the turbulent decade
of the 1990s while its matched pair was
less successful. The second study followed
a similar design, matching six Chinese
startups that adapted successfully to
shifting environmental conditions with
similar, but less successful, ventures (with
an additional two companies analyzed
without a matched pair). In both studies,
conversations within the company were a
central research focus, and I conducted
hundreds of one-on-one interviews, sup-
plemented by a review of archival records
within the businesses and direct observa-
tions of management team meetings.
The development of the strategy-loop
framework is the result of that research as
well as a general review of the existing lit-
erature on iterative processes and agility
in diverse domains, including entrepre-
neurship, military theory and software
programming. Reviewing those diverse
fields provided insight into the funda-
mental characteristics of strategy loops
that appear to be robust across domains.
In addition, a review of the existing litera-
ture on decision making helped refine
and enhance the different steps within
the strategy loop.
Concepts of the strategy loop have
been implemented at several companies
to enhance their quality of discussions.
In this research, I typically worked with
dozens (and even up to hundreds) of
managers at various levels within each
firm over the span of months (and some-
times years). A typical session would last
approximately two days, during which
the participants would learn about the
framework, complete a diagnostic exer-
cise to identify common breakdowns in
their organization at each stage in the
strategy loop and collectively brainstorm
to develop tactics for overcoming those
obstacles. I have also worked directly
with various management teams, observ-
ing their formal and informal meetings
to identify impediments to effective dis-
cussions and to coach the participants
on potential ways to improve those dis-
cussions.
About the Research
i. D.N. Sull and M. Escobari, “Success Against the Odds: What Brazilian Champions Teach Us About Thriving in Unpredictable Markets” (São Paulo: Elsevier, 2003); and D.N. Sull with
Y. Wang, “Made in China: What Western Managers Can Learn From Trailblazing Chinese Entrepreneurs” (Boston: Harvard Business School Press, 2003).
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 33SLOANREVIEW.MIT.EdU
leader, lapse into “groupthink” or ignore the available data
when forming conclusions. One warning sign is when some
participants check out of the conversation altogether, perhaps
because they believe the leader has already made a decision in a
“meeting before the meeting” and is just trying to obtain every-
one’s buy-in.
One of the most dangerous pitfalls is when a team prema-
turely develops a “bias for action.” This risk is particularly
acute among managers who pride themselves on getting the
job done. The result: The team shortchanges the sense-making
discussion and jumps right into a debate about what to do and
how to do it. But if the conversation rushes too quickly through
the messy thrashing around of sense making, managers risk
diving into the details of implementation before they’ve ex-
plored alternative assessments, surfaced and checked key
assumptions, or tested the fit between their interpretation and
the facts on the ground. Executives can mitigate this risk by
separating discussions to make sense from those to make
choices. For example, the top management team of Diageo
Ireland, which handles alcoholic beverages such as Guinness,
Smirnoff and Baileys, breaks the monthly performance man-
agement process into distinct meetings. On the second day of
the month, managers update their assessment of the market
situation and identify possible issues, and on the seventh day
they decide what to do, thereby reducing the risk of short-
changing sense making in a rush to action. When action
proposals do arise in sense-making discussions, the leader can
dig backward to unearth and examine the assumptions that
underlie the plan of action rather than rush forward into de-
tails of implementation. Questions that help uncover that
information include, “If that’s the solution, what exactly is the
problem?” and “What fresh data would convince us that this is
the wrong course of action?”7
Guiding discussions to make sense requires a distinct set of
management traits. The first is coup d’oeil, or the ability to grasp
the essence of a situation based on limited data, akin to a person
quickly being able to visualize the overall picture of a jigsaw
puzzle after glimpsing just a few pieces. Another critical attribute
is curiosity. Managers with that trait remain open to new inter-
pretations and are likely to explore unfamiliar ways of framing a
situation. Curiosity also helps people remain alert to weak signals
from many different sources — an important skill because the
crucial piece of a puzzle often comes from an unexpected source.
That’s why some leaders use specific techniques to reinforce their
curiosity. Robert Rubin, the former U.S. Treasury secretary and
co-managing director of Goldman Sachs & Co., would tackle any
new situation, from evaluating a risk arbitrage deal to managing
an economic crisis, by pulling out a pad of yellow legal paper to
write down a long list of questions — in stark contrast to many
managers who try to affirm authority by asserting answers rather
than asking good questions.8 Finally, leaders need to do more
than just tolerate different points of view; they must actively seek
them out. Various tactics can help, including arguing the oppo-
site of a given position and appointing a devil’s advocate to probe
contrary views.
Making Choices
The objective of discussions to make choices is a small set of
clear priorities that will focus organizational resources and at-
tention. Determining the right priorities is a critical function of
management under any circumstances, but the process is all the
more important (and difficult) in dynamic markets. In such
environments, the constant deluge of potential opportunities
In novel situations, the best interpretation is rarely obvious, and the obvious one is often wrong. Therefore,
the discussion leader must ensure that participants feel safe to put forth alternative interpretations.
The strategy loop is an iterative process that consists of four
major steps.
TheStrategyLoop
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU
Make Things
Happen
Make
Revisions Make Sense
Make
Choices

Strategy
34 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007 SLOANREVIEW.MIT.EdU
and threats can lead managers to hedge their bets against every
foreseeable contingency, thereby spreading corporate resources
too thin and failing to execute on key initiatives. Conversations
to make choices conclude when a group agrees on a set of pri-
orities that are both consistent with its interpretation of the
situation and sufficiently concrete to be understood by everyone
required to execute the strategy.
These conversations are, at their core, about making hard
trade-offs. As a result, the leader should establish a tone of re-
spectful argumentation, in which team members can express
valid disagreements that might otherwise simmer below the
surface. Without active efforts to stimulate debate, these conver-
sations can easily drift toward superficial agreement while
unresolved conflicts lurk below. The danger is particularly acute
if prioritization becomes politicized — participants from differ-
ent business units refrain from challenging an initiative to
respect a colleague’s turf, for example, or they horse-trade sup-
port for their pet initiatives. Leaders can actively counterbalance
such tendencies by insisting that all choices be made in public
meetings, thus adding transparency to the process, which can
help keep political prioritization in check. In addition, holding
team members collectively responsible for delivering on priori-
ties will increase their willingness to raise potential conflicts in
order to avoid being held accountable for initiatives that are ill-
advised for the organization as a whole.
Discussions to make choices frequently derail when people
add priorities without either increasing resources or removing
other initiatives. Such cases of priority proliferation can arise
when managers make decisions by focusing on specific issues in
isolation without considering the existing portfolio of activities
going on within the organization. As a result, decision makers fail
to consider which current activities they should terminate to free
the resources required for a new initiative and, over time, the re-
sult is a plethora of so-called priorities.
To avoid priority proliferation, managers can inject discipline
into the prioritization process by making choices more explicitly
and systematically. At Diageo Ireland, for instance, issues are tri-
aged into one of three categories: soft opportunities or threats,
which receive ongoing monitoring but no action; hard opportu-
nities or threats, which require immediate action and become a
priority within the company; and nonissues, which are dropped
from the agenda. Teams can also adopt a small set of simple rules
to guide the prioritization process.9 Consider All America Latina
Logistica S.A., which began life as a privatized branch of Brazil’s
freight railway. The new company had only $15 million for
capital spending to offset decades of underinvestment. So, to
select from among countless capital budgeting proposals, man-
agement adopted a set of simple rules, such as “eliminate
bottlenecks to growing revenues,” “lowest up-front cash beats
highest net present value” and “reuse of existing resources beats
acquiring new.”
Simple rules can also help prevent discussions from bogging
down in an endless quest for perfect agreement. Achieving
consensus is, of course, desirable, but the process takes time,
and the costs of delay can often outweigh the benefits, particu-
larly in fast-moving markets. Indeed, research on successful
decision making in such environments has found that the most
successful companies did not seek complete consensus, but
neither did they go to the other extreme of having one person
call all the shots.10 Instead, they followed a policy of “qualified
consensus,” in which the top management team would seek
agreement up to a certain point and then invoke a set of pre-
specified rules. The rules depended on the team and the
decision; for example, the person with the most authority (or
functional expertise) might decide, or the team might take a
vote. Interestingly, the exact rules mattered less than the fact
that they were clear, considered to be legitimate and known by
everyone in advance.
In discussions to make choices, the central leadership quality
is decisiveness, and a related trait is the ability to say no. Gener-
ally speaking, the hardest choices are not about deciding what to
do; instead, they involve determining what not to do (or what to
stop doing). Because such decisions might be unpopular, they
must typically be based on a compelling rationale grounded in
the overarching strategy and objectives of the organization. In
making choices, managers should also consider the overall en-
terprise rather than setting parochial priorities that make sense
only for their individual units. Finally, leaders need sufficient
credibility to have their decisions stick. The return of company
founders Steve Jobs (Apple Inc.), Charles Schwab (Charles
Schwab Corp.) and Michael Dell (Dell Inc.) might stem in part
from the credibility they possess within their organizations,
which enables them to bring people along even when they make
very difficult decisions.
The most successful companies did not seek complete consensus, but neither did they go to the other
extreme of having one person call all the shots. Instead, they followed a policy of “qualified consensus.”
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 35SLOANREVIEW.MIT.EdU
Making Things Happen
A simple but powerful mechanism — the promise — can help
managers make things happen.11 A promise is a personal pledge
a provider makes to satisfy the concerns of a customer within or
outside the organization. Both “customer” and “provider” refer to
roles (and not individuals), which can vary depending on the
specific situation. A business unit head within a bank, for exam-
ple, is a customer when requesting technology support from the
chief technology officer. But she is a provider when supplying
products to another division.
Companies can use promises to ensure that employees un-
derstand what they need to do and that those individuals deliver
on their commitments. To a large extent, then, execution will
hinge on the quality of promises made and on the consistency
with which those commitments are honored. In this context, the
objective of the discussions to make it happen should solicit
personal promises (between employees and their managers) to
perform actions that are aligned with agreed-on priorities. The
promises might take place within an existing procedure, such as
performance management, or in off-line negotiations, but their
purpose is the same — to weave a web of commitments that
ensure coordinated action.
A common mistake is that people often equate a promise with
a contract and focus on the specific clauses of what the provider
has committed to deliver. But the conversations that lead to a
promise and keep it alive are far more important than the actual
terms of the deal. When leading such discussions, managers
should adopt a tone of supportive discipline, demanding explicit
promises and holding people accountable for them but also help-
ing those individuals to deliver on their commitments. That
support can take several forms, providing, for example, addi-
tional resources, relief from other priorities or the political cover
needed to deliver on the commitments.
Managers should remember that the most effective promises
share five fundamental characteristics: They are public, actively
negotiated, voluntary, explicit and linked to corporate priorities.
A commitment can easily derail when any of the five is absent.
For instance, private (and not public) side deals can allow people
to wriggle out of what they said they’d do. Passive (and not ac-
tive) promises occur when people agree to do something without
probing to understand what they are really signing up for. Co-
erced (and not voluntary) commitments arise when people feel
compelled to accept a request — even one that is unrealistic —
because it comes from someone more powerful in the organization.
Vague (and not explicit) commitments offer too much scope for
interpretation of what constitutes execution, making it difficult
to hold people accountable. Lastly, commitments that are ad hoc
(and not linked to corporate priorities) arise when people make
promises that might be optimal locally but are poorly aligned
with the organization’s objectives.
Scrum, which takes its name from a play in rugby, is an ap-
proach used in the software industry that exemplifies how to
elicit good promises. In the process, a programming team con-
venes in the same place and time each workday to make and track
each member’s promises publicly. During a meeting, the partici-
pants (typically fewer than 10) stand in a circle and answer the
same three questions: What have you done since the last scrum?
What will you do between now and the next scrum? And what’s
getting in the way of you delivering on your promises? The public
forum is effective because of peer pressure — people don’t want
to let down their team, nor do they want their reputations to suf-
fer from a failure to do what they said they would. Scrums also
allow the programmers to actively talk through what they are
promising, which helps ensure that the promises are sufficiently
Understanding the four types of discussions that make up
the strategy loop is necessary but not sufficient. Leaders
must also exercise judgment in deciding which type of con-
versation to have, when to have it and how to lead it most
effectively. (“Discussions Through the Strategy Loop” sum-
marizes some key differences among the four types of
discussions.) The following questions should help:
• What are we talking about? This simple question often
reveals a disturbing lack of focus in discussions.
• Are the right people in the room? Discussions to make
sense work best when different points of view are brought to
bear; making things happen requires the presence of the
people who will ultimately do the work; and discussions for
revision often benefit from an outside viewpoint.
• Are we currently talking about the right thing? Manag-
ers must make a call on what conversation is appropriate for
the current situation. Are people jumping to choices before
they’ve made sense of what is going on, for example, or are
they revisiting assumptions when they should instead be get-
ting things done? The timing of when to shift a conversation
from one stage to another is a crucial decision that execu-
tives must make.
• Does the conversation have the right tone? Managers
must understand what an effective discussion sounds like for
each step of the strategy loop. For example, they should estab-
lish and maintain a spirit of open inquiry during discussions to
make sense of a situation, and they should promote respectful
arguments during discussions to make choices.
• Are we skipping key conversations? Execution-focused
teams are particularly prone to ignore discussions to make
sense and make revisions, while more strategic groups might
favor discussions about the market but omit critical discus-
sions to ensure that everyone does the necessary work.
WhatAreWeTalkingAbout?
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU

Strategy
36 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007 SLOANREVIEW.MIT.EdU
explicit for others to adjust their behaviors accordingly. More-
over, work is not assigned — instead, people volunteer for it
— and everyone’s commitments are always linked back to the
priorities set in monthly meetings with customers.
In discussions to make things happen, the most important
leadership trait is trustworthiness. Here, a manager can set the
right tone by consistently honoring his or her own promises.
When making a commitment, an executive takes on the respon-
sibility for all the unexpected contingencies that could occur. As
such, overcoming inevitable setbacks and obstacles requires flex-
ible tenacity — the ability to try different courses of action until
the desired results are achieved. Finally, a leader must inspire oth-
ers to make ambitious promises without coercing them to do so,
and one of the most effective ways to accomplish that is by link-
ing the assignment to a mission or corporate objective that
matters to the person making the commitment.
Making Revisions
Managers need to recognize emerging patterns in order to an-
ticipate new opportunities and threats. But spotting such
patterns also requires people to revise and sometimes even
abandon their existing mental models, and therein lies the rub.
When a person’s established patterns of thinking clash with
changing circumstances, the existing mental models typically
prevail. But letting go of the old is as important as spotting the
new. Thus managers must keep their mental models fluid,
modifying them in light of changes in the broader context. And
they must remain open to the possibility of abandoning those
established models altogether.
In any discussion to make revisions, people should treat ac-
tions as experiments: They should analyze what’s happened
and use the results to revise their assumptions, priorities and
promises. As such, the appropriate time to have such conversa-
Discussions at each stage in a strategy loop have different objectives, face different pitfalls and require distinct management
approaches to improve the quality of the conversation.
DiscussionsThroughtheStrategyLoop
MakeSense MakeChoices MakeThingsHappen MakeRevisions
Objective Develop a shared mental
model of a situation
Agree on clear priorities to
guide action and resource
allocation
Ensure that people make
good promises and deliver
Sense anomalies and
revise key assumptions
Appropriate
Tone
Open inquiry Respectful argumentation Supportive discipline Dispassionate
analysis
Information
Support
Shared dashboard of
real-time, granular data
Ongoing monitoring of
“hard” and “soft” priorities
Monitor performance
against promises
Variance reporting of
key variables to spot
anomalies
Required
Leadership
Traits
• Coup d’oeil
• Curiosity
• Empathy to see other
points of view
• Decisiveness
• Enterprise perspective
• Credibility to make the
call
•Trustworthiness
• Flexible tenacity
• Ability to inspire others
• Intellectual
humility
• Respect for
other viewpoints
• Sensitivity to anomalies
Pitfalls • Advocating pre-existing
positions
• Anchoring too quickly on
one viewpoint
• Bias for premature
action
• Superficial agreement
• Politicized prioritization
• Priority proliferation
• Searching for complete
consensus
• Private promises
• Passive agreement
• Meaningless yes
• Implicit agreements
• What without why
• Blame game
• Escalating commitment
to failed course of action
• Cognitive biases toward
confirming evidence
HelpfulTips • Question assumptions
• Interact frequently
• Explicit prioritization
• Simple rules to prioritize
• Publicly monitor
promises
• Link promises to
priorities
• Build in regular reviews
• Bring in external
reviewers
Killer
Questions
• What fresh data would
convince us that our
assessment is wrong?
• What will we stop doing? • What did you promise
to do?
• What have you done?
• What is hindering you?
• What did we expect to
happen versus what
really happened?
• Why the difference?
• What should we change?
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 37SLOANREVIEW.MIT.EdU
tions is after the team has reached a significant milestone in
making things happen. Discussions to make revisions tend to
add the most value after a prolonged period of heads-down
execution, when team members have not yet had the time to
determine whether the results have confirmed their original
assessment of the situation. Shifts in contextual factors, such as
regulatory changes or unexpected moves by competitors, al-
most always create a gap between initial assumptions and how
events actually unfold. Consequently, leaders need to design in
regular occasions for people to pause and reflect on what the
team has learned.
In principle, discussions to make revisions are simple: The
team discusses what it expected to happen (and why) versus
what actually happened, and then it explores any gaps between
expectations and reality. Leaders can facilitate these discus-
sions by explicitly framing assumptions as hypotheses and
actions as experiments.12 But even when the conversations are
presented clearly as retrospective opportunities to learn, they
remain tricky and delicate. People might still feel threatened by
the prospect of having their actions scrutinized and criticized,
and they could personalize feedback as a negative reflection on
their competence, judgment or motivation. To avoid that, the
discussions should maintain a tone of dispassionate analysis
— think of a scientist in a lab coat objectively evaluating re-
sults from an experiment.
The fear of blame isn’t the only obstacle. Psychologists have
documented a depressingly long list of factors that keep people
locked into the confines of their established mental models. For
example, people often escalate their commitment to a failed
course of action in order to avoid admitting any mistakes, or they
fixate on data that confirm their expectations while ignoring or
downplaying any contradictory information.
Given all the obstacles, organizations must go out of their
way to incorporate frequent and rigorous opportunities for re-
vision into their strategy loops. Venture capital firms, for
example, typically stage their funding in rounds, which forces
the partners and entrepreneurs to reexamine a startup’s perfor-
mance against its business plan and to consider shifts in the
market and other changes in circumstances. In fact, many ven-
ture capitalists view their most important role as that of
protecting their partners from falling in love with a bad invest-
ment. So they regularly engage in hard-hitting and skeptical
evaluations of one another’s deals, asking questions along the
lines of, “If this company walked in the door today, would we
invest?” and “Why shouldn’t we cut our losses right now?” The
partners of Onset Venture Services Corp., an early-stage venture
capital firm based in Menlo Park, California, have gone even
further by instituting a simple rule: They don’t scale a startup
until its business model has morphed at least once, building in
the expectation that not only is it OK to adjust the model, it’s
required. Consequently, Onset selects entrepreneurs to fund in
large part based on their ability to learn and adapt to shifting
circumstances. Moreover, to inject a more objective perspective
into the process, the Onset partners invite later-stage venture
capitalists from the outside to help them evaluate the progress
and prospects of their portfolio companies.
The fundamental leadership trait required for revision is
intellectual humility, which is admittedly not the most com-
mon attribute among executives. But in an uncertain world,
managers must acknowledge that their mental models are
merely simplified maps of complex terrain based on provi-
sional knowledge that is subject to revision in light of new
information. That humility can help executives to actively seek
out disconfirming information that exposes inaccuracies in
their maps. On a related note, executives must have respect for
other points of view — not just because it will smooth the road
for implementation and is desirable in and of itself, but also
because it will increase the likelihood that they will hear and
consider alternative perspectives that might lead to a revision
of past assumptions. Finally, managers should remain alert to
any new information that doesn’t jibe with their expectations.
Many anomalies provide clues to outdated and otherwise inac-
curate assumptions, and people who discover and act on that
information can seize the initiative from rivals who are slower
to respond. When managers observe an anomaly, they should
investigate it firsthand until they’re satisfied that they under-
stand the source of the discrepancy.
some industries — heavily regulated utilities, for example
— do not often produce new information that would challenge
a company’s strategy. In such stable contexts, the traditional
linear approach to strategy might suffice. But most markets
frequently generate high levels of strategically relevant informa-
tion. In such industries — call them volatile, unpredictable,
Many anomalies provide clues to outdated and otherwise inaccurate assumptions, and people who
discover and act on that information can seize the initiative from rivals who are slower to respond.
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU

Strategy
38 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007
turbulent, high-velocity, hypercompetitive, chaotic or uncertain
— the complex interactions of multiple variables (geopolitics,
technical innovation, capital market swings, competitive dy-
namics, shifting consumer preferences and so on) influence a
company’s best course of action and ultimate performance.13
Each of these variables is individually uncertain, and their
myriad potential interactions fundamentally defy prediction.
These dynamic markets throw out a steady stream of opportu-
nities and threats, and managers can neither predict nor control
the form, magnitude or timing of future events with accuracy.
In such environments, companies succeed to the extent that
they can respond to shifting circumstances. Strategy loops, with
their inherent ability to incorporate and translate new informa-
tion into action, provide an effective framework for organizations
to close the gap between strategy and execution. Managers who
master the strategy loop’s four types of discussions will be able
to spot emerging opportunities, seize them and make mid-
course corrections more effectively than others who stumble
through those steps.
REFERENCES
1. For a review of the escalation of commitment literature, see J. Brock-
ner, “The Escalation of Commitment to a Failing Course of Action:
Toward Theoretical Progress,” Academy of Management Review 17,
no. 1 (January 1992): 39–61.
2. See P. Ghemawat, “Commitment: The Dynamic of Strategy” (New
York: Free Press, 1991). Ghemawat argues that strategy consists of
making commitments or infrequent large changes in resources that have
large and enduring effects on a company’s future alternatives. The im-
portance of these decisions implies that managers can and should
clearly analyze their consequences long into the future. Ghemawat’s ar-
gument hinges on the assumption that managers can identify what
matters ex ante and can analyze the consequences of their actions, al-
though he, of course, admits the presence of uncertainty.
3. See R. Amit and P.J.H. Schoemaker, “Strategic Assets and Organiza-
tional Rent,” Strategic Management Journal 14, no. 1 (January 1993):
33-46. Amit and Schoemaker acknowledge that their view offers little
guidance to managers. See also M.E. Raynor, “The Strategy Paradox:
Why Committing to Success Leads to Failure (And What to Do About It)”
(New York: Currency, 2007).
4. See A. Schlesinger Jr., “A Thousand Days: John F. Kennedy in the
White House” (Boston: Houghton Mifflin, 1965); and R. Kennedy, “Thir-
teen Days: A Memoir of the Cuban Missile Crisis” (New York: W.W.
Norton, 1969). For an excellent analysis of decision making by Kenne-
dy’s team, see M.A. Roberto, “Deciding How to Decide,” chap. 2 in “Why
Great Leaders Don’t Take Yes For an Answer” (Upper Saddle River,
New Jersey: Wharton School Publishing, 2005).
5. See K.M. Eisenhardt, “Making Fast Strategic Decisions in High-Veloc-
ity Environments,” Academy of Management Journal 32, no. 3
(September 1989): 543-576; and K.M. Eisenhardt, “Speed and Strategic
Choice: How Managers Accelerate Decision Making,” California Man-
agement Review 32, no. 3 (spring 1990): 39-54.
6. See A.C. Edmondson, “Psychological Safety and Learning Behavior
in Work Teams,” Administrative Science Quarterly 44, no. 2 (June
1999): 350-383; and A.C. Edmondson, “Speaking Up in the Operating
Room: How Team Leaders Promote Learning in Interdisciplinary Ac-
tion Teams,” Journal of Management Studies 40, no. 6 (September
2003): 1419-1452. Edmondson’s construct of psychological safety is
critical throughout the strategy cycle, but it takes a slightly different
form in each step. In making sense, for example, psychological safety
ensures that participants feel safe to broach alternative interpretations
of what is going on, while in making things happen providers should
feel secure to negotiate what they need before they can make a bind-
ing performance promise.
7. This is a slightly modified version of “Alexander’s question” described
in R.E. Neustadt and E.R. May, “Thinking in Time: The Uses of History
for Decision Makers” (New York: Free Press, 1988): 152-153.
8. See R.E. Rubin and J. Weisberg, “In an Uncertain World: Tough
Choices From Wall Street to Washington” (New York: Random House,
2003); and L. Endlich, “Goldman Sachs: The Culture of Success” (New
York: Alfred A. Knopf, 1999).
9. K.M. Eisenhardt and D.N. Sull, “Strategy As Simple Rules,” Harvard
Business Review 79 (January 2001): 107-116.
10. See Eisenhardt, “Fast Strategic Decisions,” and Eisenhardt, “Strate-
gic Choice.”
11. For an in-depth perspective on the promise-based view of the firm,
see D.N. Sull and C. Spinosa, “Promise-Based Management: The Es-
sence of Execution,” Harvard Business Review 85 (April 2007): 78-86;
and D.N. Sull and C. Spinosa, “Using Commitments to Manage Across
Units,” MIT Sloan Management Review 47, no. 1 (fall 2005): 73-81.
12. For a thoughtful and practical guide to after-action reviews based on
practices within the U.S. Army, see D.A. Garvin, “Learning in Action: A
Guide to Putting the Learning Organization to Work” (Boston: Harvard
Business School Press, 2003): 106-116.
13. See F. Emery and E. Trist, “The Causal Texture of Organizational En-
vironments,” Human Relations 18 (1965): 21-32. In their discussion of
uncertain markets, Emery and Trist emphasize complexity (that is, multi-
ple factors that influence performance) and dynamism (that is, the rate
of change of those variables) and imply the role of interactions.
Reprint 48412. For ordering information, see page 1.
Copyright © Massachusetts Institute of Technology, 2007. All rights reserved.
Managers who master the four types of discussions will be able to spot emerging opportunities, seize
them and make midcourse corrections more effectively than others who stumble through those steps.
SLOANREVIEW.MIT.EdUSLOANREVIEW.MIT.EdU
Reprint 48412.
Copyright © Massachusetts Institute of Technology, 2007. All rights reserved.

PDFs ■ Reprints ■ Permission to Copy ■ Back Issues
Electronic copies of MIT Sloan Management Review
articles as well as traditional reprints and back issues can
be purchased on our Web site: sloanreview.mit.edu or
you may order through our Business Service Center
(9 a.m.-5 p.m. ET) at the phone numbers listed below.
To reproduce or transmit one or more MIT Sloan
Management Review articles by electronic or mechanical
means
(including photocopying or archiving in any
information storage or retrieval system)
requires written
permission.
To request permission, use our Web site
(sloanreview.mit.edu), call or e-mail:
Toll-free
in U.S. and Canada: 877-727-7170
International: 617-253-7170
Fax: 617-258-9739
e-mail: smrpermissions@mit.edu
MIT Sloan Management Review
77 Massachusetts Ave., E60-100
Cambridge, MA 02139-4307
e-mail: smrorders@mit.edu