WhenTeamscantDecide.pdf
hbr.org | November 2008 | Harvard Business Review 121
THE EXECUTIVE TEAM is deliberating about a critical stra-
tegic choice, but no matter how much time and effort the
team members expend, they cannot reach a satisfactory
decision. Then comes that uncomfortable moment when
all eyes turn to the CEO. The team waits for the boss to
make the fi nal call, yet when it’s made, few people like the
decision. Blame, though unspoken, is plentiful. The CEO
blames the executives for indecisiveness; they resent the
CEO for acting like a dictator. If this sounds familiar, you’ve
experienced what I call the dictator-by-default syndrome.
For decades this dynamic has been diagnosed as a prob-
lem of leadership or teamwork or both. To combat it, com-
panies use team-building and communications exercises
that teach executives how to have assertive conversations,
give and receive feedback, and establish mutual trust. In
doing so, they miss the real problem, which lies not with
the people but with the process. This sort of impasse is
inherent in the act of arriving at a collective preference on
the basis of individual preferences. Once leadership teams
understand that voting-system mathematics are the cul-
prit, they can stop wasting time on irrelevant psychological
exercises and instead adopt practical measures designed
to break the impasse. These measures, proven effective in
scores of strategy off-sites for companies of all sizes, enable
teams to move beyond the blame cycle to a no-fault style
of decision making.
When Teams Can’t Decide Are stalemates on your leadershipteam making you a dictator by default? Stop blaming your people – start fi xing the process.
Best PracticeBY BOB FRISCH
Jen
Hsi
eh
1592 Frisch.indd 1211592 Frisch.indd 121 10/6/08 1:02:30 PM10/6/08 1:02:30 PM
122 Harvard Business Review | November 2008 | hbr.org
Best Practice When Teams Can’t Decide
Asking the ImpossibleReaching collective decisions based on
individual preferences is an imperfect
science. Majority wishes can clash when
a group of three or more people at-
tempts to set priorities among three or
more items. This “voting paradox,” fi rst
noted in the eighteenth century by the
Marquis de Condorcet, a French math-
ematician and social theorist, arises
because different subsets of the group
can generate confl icting majorities for
all possible alternatives (see the exhibit
“The Boss Is Always Wrong”). A century
and a half later, renowned economist
Ken Arrow developed his impossibil-
ity theorem, which established a series
of mathematical proofs based on Con-
dorcet’s work.
Suppose a nine-person leadership
team that wants to cut costs is weighing
three options: (a) closing plants, (b) mov-
ing from a direct sales force to distribu-
tors, and (c) reducing benefi ts and pay.
While any individual executive may be
able to “rack and stack” her preferences,
it’s possible for a majority to be simulta-
neously found for each alternative. Five
members might prefer “closing plants”
to “moving sales to distributors” (a > b),
and a different set of fi ve might prefer
“moving sales” to “reducing benefi ts and
pay” (b > c). By the transitive property,
“closing plants” should be preferred to
“reducing benefi ts and pay” (a > c). But
the paradox is that fi ve members could
rank “reducing benefi ts and pay” over
“closing plants” (c > a). Instead of being
transitive, the preferences are circular.
When the CEO is fi nally forced to
choose an option, only a minority of
team members will agree with the de-
cision. No matter which option is se-
lected, it’s likely that different majorities
will prefer alternative outcomes. More-
over, as Arrow demonstrated, no voting
method – not allocation of points to al-
ternatives, not rank-ordering of choices,
nothing – can solve the problem. It can
be circumvented but not cured.
Although the concept is well under-
stood in political science and economics
and among some organizational theo-
rists, it hasn’t yet crossed over to prac-
tical management. Understanding this
paradox could greatly alter the way ex-
ecutive teams make decisions.
Acknowledging the ProblemTo circumvent the dictator-by-default
syndrome, CEOs and their teams must
fi rst understand the conditions that give
rise to it. The syndrome is perhaps most
obvious at executive off-sites, but it can
crop up in any executive committee
meeting of substance.
Most executive teams are, in effect,
legislatures. With the exception of the
CEO, each member represents a signifi –
cant constituency in the organization,
from marketing to operations to fi nance.
No matter how many times a CEO asks
team members to take off their func-
tional hats and view the organization
holistically, the executives fi nd it diffi cult
to divorce themselves from their func-
tional responsibilities. Because the team
often focuses on assigning resources and
setting priorities, members vie for allo-
cations and approval for favored proj-
ects. When more than two options are
on the table, the scene is set for the CEO
to become a dictator by default.
More insidiously, the problem exists
even when a team is considering an
either/or choice, despite the fact that
the voting paradox requires three or
more options. Framing strategy consid-
erations as binary choices – “We must
either aggressively enter this market
or get out of this line of business alto-
gether” – appears to avert the problem.
However, such choices always include
a third, implied alternative: “Neither of
the above.” In other words, there could
be circular majorities for entering the
market, for exiting the business, and for
doing neither.
Take, for example, the ubiquitous
business case, which usually offers a
single, affirmative recommendation:
“We should aggressively enter this mar-
ket now.” The only apparent alternative
is to forgo the market – but some team
members may want to enter it more ten-
tatively, others may want to enter an ad-
jacent market, and still others may want
to defer the decision until the market
potential becomes clearer.
The use of the business case, which
forces decisions into a yes-or-no frame-
work, is a tacit admission that groups are
not good at discussing and prioritizing
multiple options. Further, when a team
of analysts has spent six months work-
ing up the business case and only a half
hour has been allotted to the item on
the agenda, dissenting team members
may be reluctant to speak up. Questions
from the heads of sales and marketing,
who have spent only a day or two with a
briefi ng book and 20 minutes watching
a PowerPoint presentation, would most
likely be treated as comments tossed
from the peanut gallery. So the team
remains silent and unwittingly locked
in the voting paradox. Ultimately, in or-
der to move on to the next agenda item,
either the team appears to reach a
majority view or the CEO issues a fi at.
In reality, however, there may be com-
peting opinions, alternative majority
opinions, and dissatisfaction with the
outcome – all of them unstated.
Managing the ImpossibleOnce CEOs and their teams under-
stand why they have trouble making
decisions, they can adopt some straight-
IDEA IN BRIEF
When executive teams hit an ■
impasse deliberating on an important decision, they often look to the CEO to make the fi nal call, only to be displeased with the outcome.
The CEO blames the team ■
for indecisiveness; the team resents the CEO for acting like a dictator.
This problem arises because ■
groups try to reach consen-sus on the basis of individual preferences.
Use the tactics described here ■
to circumvent this dictator-by-default syndrome and create genuine team alignment.
1592 Frisch.indd 1221592 Frisch.indd 122 10/6/08 1:02:39 PM10/6/08 1:02:39 PM
hbr.org | November 2008 | Harvard Business Review 123
forward tactics to minimize potential
dysfunction.
Articulate clearly what outcome you are seeking. It’s surprising how of-
ten executives assume that they are talk-
ing about the same thing when in fact
they are talking past one another. In a
discussion of growth, for instance, some
may be referring to revenue, others to
market share, and others to net income.
The discussion should begin with agree-
ment on what outcome the team is try-
ing to achieve. If it’s growth, then do all
the members agree on which measures
are most relevant?
In the absence of clearly articulated
goals, participants will choose options
based on unspoken, often widely differ-
ing, premises, creating a situation that is
ripe for the dictator-by-default syndrome.
One division of a major industrial com-
pany, for example, was running out of
manufacturing capacity for a commod-
ity product made in the United States
and a specialty product made in Western
Europe. Because costs of labor and raw
materials were high in both places, the
leadership team was considering what
seemed like an obvious choice: shutting
down the U.S. plant and building a plant
in China, where costs were lower and
raw materials were closer, to handle the
commodity business and any growth in
the specialty business. Most members of
the team assumed that the desired out-
come was to achieve the highest possible
return on net assets, which the move to
China might well have accomplished.
However, the CEO had been in dis-
cussions with corporate managers who
were primarily concerned with alloca-
tion of overhead throughout the enter-
prise. The move to China would mean
shutting down an additional plant that
supplied raw materials to the U.S. plant,
with implications for corporate earnings.
Once the division team fully understood
what outcome the parent company
desired – to minimize overhead costs
without taking a hit on earnings – it
could work on solving the capacity prob-
lem in a way that honored the parent’s
strictures.
It’s essential to keep discussion of the
desired outcome distinct from discus-
sion about how to achieve it. Sometimes,
simply articulating the desired outcome
will forestall or dissolve disagreement
about solutions because the options can
be tested against an accepted premise.
It may also help avert the political horse
trading that can occur when executives
try to protect their interests rather than
aiming for a common goal.
Provide a range of options for achieving outcomes. Once the team at
the industrial company had articulated
the desired outcome, it could break the
simplistic “accept,” “reject,” and “defer”
alternatives into a more nuanced range
of options: build a specialty plant in
China; beef up the plant in Western
Europe; or build a commodity plant in
China and gradually decommission the
U.S. plant.
Test fences and walls. When teams
are invited to think about options, they
almost immediately focus on what they
can’t do – especially at the divisional
level, where they may feel hemmed in
by corporate policies, real or imagined.
Often the entire team not only assumes
that a constraint is real but also shies
away when the discussion comes any-
where near it. When team members cite
a presumed boundary, my colleagues
and I encourage them to ask whether
it’s a wall, which can’t be moved, or a
fence, which can.
For example, one division of a glo-
bal provider of fi nancial services was
A management team is attempting to select a fl eet vehicle for its com-pany’s senior executives. When asked to rank three choices – BMW, Lexus, and Mercedes – the individual team mem- bers reach an impasse.
To break it, the CEO intervenes and picks BMW. But as the table shows, two-thirds of the team would have preferred a Lexus. Had he chosen Lexus, however, two-thirds of the team would have preferred
Mercedes. And had he chosen Mercedes, two-thirds of the team would have preferred BMW. Instead of being transi-tive – Lexus beats BMW; Mercedes beats Lexus; therefore Mercedes beats BMW – the choice is circular. Whatever decision the boss makes, the majority of his team is rooting for a different option. Unjustly, but not surprisingly, he is considered a dictator.
THE VOTING PARADOX
The Boss Is Always Wrong
It’s essential to keep discussion of the desired outcome distinct from discussion about how to achieve it.
First Choice Second Choice Third Choice
Lou BMW Mercedes Lexus
Sue Mercedes Lexus BMW
Stu Lexus BMW Mercedes
1592 Frisch.indd 1231592 Frisch.indd 123 10/6/08 1:02:46 PM10/6/08 1:02:46 PM
124 Harvard Business Review | November 2008 | hbr.org
Best Practice When Teams Can’t Decide
looking at new avenues for growth. Al-
though expanding the division’s offer-
ings to include banking services was
a promising possibility, the executive
team never considered it, assuming
that corporate policy prohibited the
company from entering banking. When
the division head explicitly tested that
assumption with her boss, she found
that the real prohibition – the wall –
was against doing anything that would
bring certain types of new regulatory
requirements. With that knowledge, the
division’s executive team was able to
develop strategic options that included
some features of banking but avoided
any new regulations.
Surface preferences early. Like ju-
ries, executive teams can get an initial
sense of where they stand by taking
nonbinding votes early in the discus-
sion. They can also conduct surveys in
advance of meetings in order to identify
areas of agreement and disagreement as
well as the potential for deadlock.
A global credit card company was de-
ciding where to invest in growth. Ordi-
narily, executive team members would
have embarked on an open-ended dis-
cussion in which numerous countries
would be under consideration; that tac-
tic would have invited the possibility of
multiple majorities. Instead, they con-
ducted a straw poll, quickly eliminating
the countries that attracted no votes and
focusing their subsequent discussion on
the two places where there was the most
agreement.
Using weighted preferences is another
way to narrow the decision-making
fi eld and help prevent the dictator-by-
default syndrome. The life and annuities
division of a major insurance company
had developed a business plan that in-
cluded a growth in profi t of $360 mil-
lion. The executive team was trying to
determine which line of business would
deliver that growth. Instead of casting
equally weighted votes for various lines
of business, each executive was given
poker chips representing $360 million
and a grid with squares representing the
company’s products and channels. Team
members distributed their chips accord-
ing to where they thought the projected
growth was likely to be found. After
discussing the results they repeated the
exercise, fi nding that some agreement
emerged.
By the third and fi nal round of the ex-
ercise, this weighted voting had helped
them narrow their discussion to a hand-
ful of businesses and channels, and gen-
uine alignment began to develop among
team members. Equally weighted votes
might have locked the executive team
into the voting paradox, but this tech-
nique dissolved the false equality of al-
ternatives that is often at the root of the
problem. Proposing options early and
allowing people to tailor them reduces
the likelihood that executives will be
forced into a stalemate that the CEO
has to break.
State each option’s pros and cons. Rather than engaging in exercises about
giving feedback or learning how to have
assertive conversations, executives can
better spend their time making sure
that both sides of every option are force-
fully voiced. That may require a devil’s
advocate.
The concept of a devil’s advocate orig-
inated in the Roman Catholic Church’s
canonization process, in which a lawyer
Proposing options early and allowing people to tailor them reduces the likelihood of a stalemate.
1592 Frisch.indd 1241592 Frisch.indd 124 10/6/08 1:02:55 PM10/6/08 1:02:55 PM
is appointed to argue against the can-
onization of a candidate – even the most
apparently saintly. Similarly, in law,
each side fi les its own brief; the defense
doesn’t simply respond off-the-cuff to
the plaintiff’s argument.
In business, however, an advocate for
a particular option typically delivers a
presentation that may contain some
discussion of risk but remains entirely
the work of someone who is sold on the
idea. Members of the executive team
are expected to agree with the business
case or attack it, although they may have
seen it only a few days before the meet-
ing and thus have no way of producing
an equally detailed rebuttal or offering
solid alternatives. Further, attacking the
business case is often perceived as at-
tacking the person who is presenting it.
Frequently the only executives with
open license to ask tough, probing ques-
tions are the CEO and the CFO, but even
they lack the detailed knowledge of the
team advocating the business case.
By breaking the false binary of a busi-
ness case into several explicit and im-
plicit alternatives and assigning a devil’s
advocate to critique each option, you
can depersonalize the discussion, mak-
ing thorough and dispassionate counter-
arguments an expected part of strategic
deliberations. This approach is espe-
cially valuable when the preferences of
the CEO or other powerful members
of the team are well known. If assign-
ing a devil’s advocate to each option
appears too cumbersome, try a sim-
pler variant: Have the CEO or a meet-
ing facilitator urge each team member
to offer two or three suggestions from
the perspective of his functional area.
Instead of unreasonably asking execu-
tives to think like a CEO, which usually
elicits silence or perfunctory comments,
this tactic puts team members on the
solid ground of their expertise and trans-
forms an unsatisfying false binary into
far more options for discussion.
A major internet entertainment com-
pany adopted a novel version of the
devil’s advocate approach. The company
maintains a council to consider its many
Change lives. Change organizations. Change the world.
1592 Frisch.indd 1251592 Frisch.indd 125 9/26/08 6:35:46 PM9/26/08 6:35:46 PM
126 Harvard Business Review | November 2008 | hbr.org
Best Practice When Teams Can’t Decide
potential investments, from upgrading
its server farms to adopting new tech-
nology to creating special entertainment
events on the web. In the past, each op-
portunity was presented to the council
as a business case by an advocate of the
investment, and each case was evaluated
in isolation.
Frustrated with this haphazard ap-
proach, the company established a new
system: The council now considers all in-
vestment proposals as a portfolio at its
monthly strategy meetings. All propos-
als follow an identical template, allow-
ing for easy comparison and a uniform
scoring system. Finally, each one needs
sign-off from an independent executive.
This system incorporates the devil’s
advocate role at two levels. For each
proposal the validating executive, not
wishing to be accountable for ground-
less optimism, considers carefully all
of the counterarguments, does a reality
check, and makes sure the sponsor ad-
justs the score accordingly. At the port-
folio level, the comparative-scoring system
reminds the team that the proposals are
competing for limited resources, which
prompts a more critical assessment.
Devise new options that preserve the best features of existing ones. Despite a team’s best efforts, executives
can still fi nd themselves at an impasse.
That is a measure of both the weighti-
ness of some strategic decisions and the
intractability of the voting paradox –
it’s not necessarily an index of executive
dysfunction.
Teams should continue to reframe
their options in ways that preserve their
original intent, be it a higher return
on net assets or greater growth. When
they feel the impulse to shoehorn deci-
sions into an either/or framework, they
should step back and generate a broader
range of options. For instance, the execu-
tive team of the property and casualty
division of a large insurer wanted to
grow either by signifi cantly increasing
the company’s share with existing agen-
cies or by increasing the total number of
agencies that sold its products. Before
the leadership team took either path, it
needed to decide whether to offer a full
line of products or a narrow line. As a
result, team members found themselves
considering four business models: (1) full
product line, existing large agencies;
(2) narrow product line, existing large
agencies; (3) full product line, more
small agencies; and (4) narrow prod-
uct line, more small agencies. Dissatis-
fi ed with those choices, they broke the
business down into 16 value attributes,
including brand, claim service, agency
compensation, price competitiveness,
breadth of product offering, and agency-
facing technology. Some of these value
attributes might apply to all four of the
original business models; others to three
or fewer. Agent-facing technology, for ex-
ample, is typical of working with many
small agencies, because their sheer num-
bers preclude high-touch relationships
with each one.
The team then graded its company
and several competitors on each attri-
bute to fi nd competitive openings that
fi t with the division’s willingness and
ability to invest. Instead of four static
choices, it now had a much larger num-
ber of choices based on different combi-
nations of value attributes. Ultimately, it
chose to bring several lagging attributes
up to market standard, elevate others
to above-market standard, and aggres-
sively emphasize still others. This turned
out to be a far less radical redirection
than the team had originally assumed
was needed.
Two Essential Ground RulesSo far, I have outlined several tactics that
leadership teams can use to circumvent
the dictator-by-default syndrome. These
tactics can be effective whether they are
used singly or in tandem. But if teams
are to thwart this syndrome, they must
adhere to two ground rules.
Deliberate confi dentially. A secure
climate for the conversation is essen-
tial to allow team members to fl oat
trial balloons and cut deals. An ex-
ecutive who knows that her speculative
remarks about closing plants may be cir-
culated throughout the company will be
reluctant to engage in the free play of
mind that unfettered strategy discus-
sion demands. Moreover, team mem-
bers whose priorities don’t prevail in
the deliberations must be able to save
face when the meeting is over. If they
are known to have “lost” or to have re-
linquished something dear to their con-
stituents, their future effectiveness as
leaders might be undermined.
Deliberate over an appropriate time frame. All too often the agendas
for strategy off-sites contain items like
“China market strategy,” with 45 minutes
allotted for the decision. The result is a
discussion that goes nowhere or an ar-
bitrary decision by the CEO that runs
roughshod over competing majorities
for other options. When new options
are devised or existing ones unbundled,
team members need time to study them
carefully and assess the counterargu-
ments. Breaking up the discussion into
several meetings spaced widely apart
and interspersed with additional analy-
sis and research gives people a chance
to reconsider their preferences. It also
gives them time to prepare their con-
stituencies for changes that are likely to
emerge as a result of a new strategy.
• • •
Leadership and communication exer-
cises have their merits. A team can’t
make effective decisions if its members
don’t trust one another or if they fail to
listen to one another. The problem I see
most often, however, is one that simply
cannot be fi xed with the psychological
tools so often touted in management lit-
erature. If executives employ the tactics
described here, which are designed to fi x
the decision-making process, they will
have far greater success in achieving real
alignment.
Bob Frisch (rfrisch@strategicoffsites.
com) is the managing partner of the
Strategic Offsites Group in Boston and
a coauthor of “Off-Sites That Work”
(HBR June 2006).
Reprint R0811JTo order, see page 139.
1592 Frisch.indd 1261592 Frisch.indd 126 10/6/08 1:03:11 PM10/6/08 1:03:11 PM
Copyright 2008 Harvard Business Publishing. All Rights Reserved. Additional restrictionsmay apply including the use of this content as assigned course material. Please consult yourinstitution's librarian about any restrictions that might apply under the license with yourinstitution. For more information and teaching resources from Harvard Business Publishingincluding Harvard Business School Cases, eLearning products, and business simulationsplease visit hbsp.harvard.edu.