delusions of success

Delusionsof Success

HowOptimismUnderminesExecutives'

I Decisions

HARVARD BUSINESS REVIEW

by Dan Lovallo and Daniel Kahneman I N 1992, OXFORD HEALTH PLANS started to build acomplex new computer system for processing claimsand payments. From the start, the project was ham-

pered by unforeseen problems and delays. As the com-pany fell further behind schedule and budget, it strug-gled, vainly, to stem an ever rising flood of paperwork.When, on October 27, 1997, Oxford disclosed that itssystem and its accounts were in disarray, the company'sstock price dropped 63%, destroying more than $3 billionin shareholder value in a single day.

JULY 2003 57

Delusions of Success

No matterhow detailed,the business

Early in the 1980s, the United Kingdom, Germany, Italy,and Spain announced that they would work togetherto build the Eurofighter, an advanced military jet Theproject was expected to cost $20 billion, and the jet wasslated to go into service in 1997. Today, after nearly twodecades of technical glitches and unexpected expenses,the aircraft has yet to be deployed, and projected costshave more than doubled, to approximately $45 billion.

In 1996, the Union Pacific railroad bought its competi-tor Southern Pacific for $3.9 billion, cre-ating the largest rail carrier in NorthAmerica. Almost immediately, the twocompanies began to have serious diffi-culties merging their operations, leadingto snarled traffic, lost cargo, and massivedelays. As the situation got worse, and thecompany's stock price tumbled, custom-ers and shareholders sued the railroad,and it had to cut its dividend and raisenew capital to address the problems.

Debacles like these are all too commonin business. Most large capital investmentprojects come in late and over budget, never living upto expectations. More than 70% of new manufacturingplants in North America, for example, close within theirfirst decade of operation. Approximately three-quartersof mergers and acquisitions never pay off-the acquiringfirm's shareholders lose more than the acquired firm'sshareholders gain. And efforts to enter new markets fareno better; the vast majority end up being abandonedwithin a few years.

According to standard economic theory, the high fail-ure rates are simple to explain: The frequency of pooroutcomes is an unavoidable result of companies takingrational risks in uncertain situations. Entrepreneurs andmanagers know and accept the odds because the rewardsof success are sufficiently enticing. In the long run, thegains from a few successes will outweigh the losses frommany failures.

This is, to be sure, an attractive argument from theperspective of executives. It effectively relieves them ofblame for failed projects-after all, they were just takingreasonable risks. But having examined this phenom-enon from two very different points of view-a businessscholar's and a psychologist's-we have come to a differ-ent conclusion. We don't believe that the high number ofbusiness failures is best explained as the result of rational

Dan Lovallo is a senior lecturer at the Australian GraduateSchool of Management at the University of New South Walesand a former strategy specialist at McKinsey & Company.Daniel Kahneman is the Eugene Higgins Professor of Psy-chology at Princeton University in New Jersey and a profes-sor of public affairs at Princeton's Woodrow Wilson School;he received the Nobel Prize in economic sciences in 2002.

choices gone wrong. Rather, we see it as a consequence offlawed decision making. When forecasting the outcomesof risky projects, executives all too easily fall victim towhat psychologists call the plarming fallacy. In its grip,managers make decisions based on delusional optimismrather than on a rational weighting of gains, losses, andprobabilities. They overestimate benefits and underesti-mate costs. They spin scenarios of success while over-looking the potential for mistakes and miscalculations.

As a result, managers pursue initiativesthat are unlikely to come in on budget oron time-or to ever deliver the expectedreturns.

Executives'overoptimism can be tracedboth to cognitive biases-to errors in theway the mind processes information-and to organizational pressures. Thesebiases and pressures are ubiquitous, buttheir effects can be tempered. By supple-menting traditional forecasting pro-cesses, which tend to focus on a com-pany's own capabilities, experiences, and

expectations, with a simple statistical analysis of analo-gous efforts completed earlier, executives can gain amuch more accurate understanding of a project's likelyoutcome. Such an outside view, as we call it, provides areality check on the more intuitive inside view, reducingthe odds that a company will rush blindly into a disas-trous investment of money and time.

Rose-Colored GlassesMost people are highly optimistic most of the time. Re-search into human cognition has traced this overopti-mism to many sources. One ofthe most powerful is thetendency of individuals to exaggerate their own talents-to believe they are above average in their endowment ofpositive traits and abilities. Consider a survey of 1 millionstudents conducted by the College Board in the 1970s.When asked to rate themselves in comparison to theirpeers, 70% ofthe students said they were above average inleadership ability, while only 2% rated themselves belowaverage. For athletic prowess, 60% saw themselves abovethe median, 6% below. When assessing their ability to getalong with others, 60% ofthe students judged themselvesto be in the top decile, and fully 25% considered them-selves to be in the top 1%.

The inclination to exaggerate our talents is amplifiedby our tendency to misperceive the causes of certainevents. The typical pattern of such attribution errors, aspsychologists call them, is for people to take credit for pos-itive outcomes and to attribute negative outcomes to ex-ternal factors, no matter what their true cause. One studyof letters to shareholders in annual reports, for example,found that executives tend to attribute favorable out-

58 HARVARD BUSINESS REVIEW


comes to factors under their control, such astheir corporate strategy or their R&D programs.Unfavorable outcomes, by contrast, were morelikely to be attributed to uncontrollable externalfactors such as weather or inflation. Similar self-serving attributions have been found in otherstudies of annual reports and executive speeches.

We also tend to exaggerate the degree of con-trol we have over events, discounting the role ofluck. In one series of studies, participants wereasked to press a button that could illuminate ared light. The people were told that whether thelight flashed was determined by a combinationof their action and random chance. Afterward,they were asked to assess what they experienced.Most people grossly overstated the influence oftheir action in determining whether the lightflashed.

Executives and entrepreneurs seem to behighly susceptible to these biases. Studies thatcompare the actual outcomes of capital invest-ment projects, mergers and acquisitions, andmarket entries with managers'original expecta-tions for those ventures shov/ a strong tendency towardoveroptimism. An analysis of start-up ventures in a widerange of industries found, for example, that more than80% failed to achieve their market-share target. The stud-ies are backed up by observations of executives. Like otherpeople, business leaders routinely exaggerate their per-sonal abilities, particularly for iimbiguous, hard-to-measuretraits like managerial skill. Their self-confidence can leadthem to assume tbat they'll be able to avoid or easily over-come potential problems in executing a project. This mis-apprehension is further exaggerated by managers' ten-dency to take personal credit for lucl<y breaks. Think ofmergers and acquisitions, for instance. Mergers tend tocome in waves, during periods of economic expansion. Atsuch times, executives can overattribute their company'sstrong performance to their own actions and abilitiesrather than to the buoyant economy. This can, in turn,lead them to an inflated belief in their own talents. Con-sequently, many M&A decisions may be the result ofhubris, as the executives evaluating an acquisition candi-date come to believe that, with proper planning and su-perior management skills, they could make it more valu-able. Research on postmerger performance suggests that,on average, they are mistaken.

Managers are also prone to the illusion that they are incontrol. Sometimes, in fact, they will explicitly deny therole of chance in the outcome of their plans. They see riskas a challenge to be met by the exercise of skill, and theybelieve results are determined purely by their own ac-tions and those of their organizations. In their idealizedself-image, these executives are not gamblers but prudentand determined agents, who are in control of both people

and events. When it comes to making forecasts, therefore,they tend to ignore or downplay the possibility of randomor uncontrollable occurrences that may impede theirprogress toward a goal.

The cognitive biases that produce overoptimism arecompounded by the limits of human imagination. Nomatter how detailed, the business scenarios used inplanning are generally inadequate. The reason is simple:Any complex project is subject to myriad problems -fromtechnology failures to shifts in exchange rates to badweather-and it is beyond the reach ofthe human imag-ination to foresee all of them at the outset. As a result,scenario planning can seriously understate the proba-bility of things going awry. Often, for instance, managerswill establish a "most likely" scenario and then assumethat its outcome is in fact the most likely outcome. Butthat assumption can be wrong. Because the managershave not fully considered all the possible sequences ofevents that might delay or otherwise disrupt the project,they are likely to understate the overall probability ofunfavorable outcomes. Even though any one of thoseoutcomes may have only a small chance of occurring, incombination they may actually be far more likeiy tohappen than the so-called most lil̂ ely scenario.

Accentuating the PositiveIn business situations, people's native optimism is furthermagnified by two other kinds of cognitive bias-anchor-ing and competitor neglect- as well as political pressuresto emphasize the positive and downplay the negative.Let's look briefiy at each of these three phenomena.

JULY 2003 59


Optimistic onesare rewarded,an organization's

Anchoring. When executives and their subordinatesmake forecasts about a project, they typically have, as astarting point, a preliminary plan drawn up by the personor team proposing the initiative. They adjust this originalplan based on market research, financial analysis, or theirown professional judgment before arriving at decisionsabout whether and how to proceed. This intuitive andseemingly unobjectionable process has serious pitfalls,however. Because the initial plan will tend to accentuatethe positive - as a proposal, it's designed to make the casefor the project-it will skew the subsequent analysistoward overoptimism. This phenomenon is the result ofanchoring, one of the strongest andmost prevalent of cognitive biases.

In one experiment that revealedthe power of anchoring, people wereasked for the last four digits of theirSocial Security number. They werethen asked whether the number ofphysicians in Manhattan is larger orsmaller than the number formed bythose four digits. Finally, they wereasked to estimate what the number ofManhattan physicians actually is. Thecorrelation between the Social Secu-rity number and the estimate wassignificantly positive. The subjectsstarted from a random series of digitsand then insufficiently adjusted their estimate awayfrom it.

Anchoring can be especially pernicious when it comesto forecasting the cost of major capital projects. Whenexecutives set budgets for such initiatives, they build incontingency funds to cover overruns. Often, however, theyfail to put in enough. That's because they're anchored totheir original cost estimates and don't adjust them suf-ficiently to account for the likelihood of problems anddelays, not to mention expansions in the scope of theprojects. One Rand Corporation study of 44 chemical-processing plants owned by major companies like 3M,DuPont, and Texaco found that, on average, the factories'actual construction costs were more than double theinitial estimates. Furthermore, even a year after start-up,about half the plants produced at less than 75% of theirdesign capacity, with a quarter producing at less than 50%.Many of the plants had their performance expectationspermanently lowered, and the owners never realized areturn on their investments.

Competitor Neglect. One ofthe key factors influenc-ing the outcome of a business initiative is competitors'behavior. In making forecasts, however, executives tendto focus on their own company's capabilities and plansand are thus prone to neglect the potential abilities andactions of rivals. Here, again, the result is an underesti-mation ofthe potential for negative events-in this case,

price wars, overcapacity, and the like. Joe Roth, the for-mer chairman of Walt Disney Studios, expressed the prob-lem well in a 1996 interview with the Los Angeles Times:"If you only think about your own business, you think,'I've got a good story department, I've got a good mar-keting department, we're going to go out and do this.'And you don't think that everybody else is thinking thesame way."

Neglecting competitors can be particularly destruc-tive in efforts to enter new markets. When a companyidentifies a rapidly growing market well suited to itsproducts and capabilities, it will often rush to gain a

beachhead in it, investing heavily inproduction capacity and marketing.The effort is often justified by the cre-ation of attractive pro forma forecastsof financial results. But such forecastsrarely account for the fact that manyother competitors will also target themarket, convinced that they, too, havewhat it takes to succeed. As all thesecompanies invest, supply outstrips de-mand, quickly rendering the new mar-ket unprofitable. Even savvy venturecapitalists fell into this trap during therecent ill-fated internet boom.

Organizational Pressure. Everycompany has only a limited amount of

money and time to devote to new projects. Competitionfor this time and money is intense, as individuals andunits jockey to present their own proposals as being themost attractive for investment. Because forecasts are crit-ical weapons in these battles, individuals and units havebig incentives to accentuate the positive in laying outprospective outcomes. This has two ill effects. First, it en-sures that the forecasts used for planning are overopti-mistic, which, as we described in our discussion of an-choring, distorts all further analysis. Second, it raises theodds that the projects chosen for investment will be thosewith the most overoptimistic forecasts-and hence thehighest probability of disappointment.

Other organizational practices also encourage opti-mism. Senior executives tend, for instance, to stress theimportance of stretch goals for their business units. Thiscan have the salutary effect of increasing motivation, butit can also lead unit managers to further skew their fore-casts toward unrealistically rosy outcomes. (And whenthese forecasts become the basis for compensation targets,the practice can push employees to behave in danger-ously risky ways.) Organizations also actively discouragepessimism, which is often interpreted as disloyalty. Thebearers of bad news tend to become pariahs, shunned andignored by other employees. When pessimistic opinionsare suppressed, while optimistic ones are rewarded, anorganization's ability to think critically is undermined.

60 HARVARD BUSINESS REVIEW


The optimistic biases of individual employees becomemutually reinforcing, and unrealistic views ofthe futureare validated by the group.

The Outside ViewFor most of us, the tendency toward optimism is un-avoidable. And it's unlikely that companies can, or wouldeven want to, remove the organizational pressures thatpromote optimism. Still, optimism can, and should, betempered. Simply understanding the sources of overopti-mism can help planners challenge assumptions, bringin alternative perspectives, and in general take a balancedview ofthe future.

But there's also a more formal way to improve the reli-ability of forecasts. Companies can introduce into theirplanning processes an objective forecasting method thatcounteracts the personal and organizational sources ofoptimism. We'll begin our exploration of this approachwith an anecdote that illustrates both the traditionalmode of forecasting and the suggested alternative.

In 1976, one of us was involved in a project to developa curriculum for a new subject area for high schools inIsrael. The project was conducted by a small team of aca-demics and teachers. When the team had been operatingfor about a year and had some significant achievementsunder its belt, its discussions turned to the question ofhow long the project would take. Everyone on the teamwas asked to write on a slip of paper the number ofmonths that would be needed to finish the project-defined as having a complete report ready for submissionto the Ministry of Education. The estimates ranged from18 to 30 months.

One ofthe team members-a distinguished expert incurriculum development - was then posed a challenge byanother team member: "Surely, we're not the only teamto have tried to develop a curriculum where none existedbefore. Try to recall as many such projects as you can.Think of them as they were in a stage comparable to oursat present. How long did it take them at that point toreach completion?" After a long silence, the curriculumexpert saad, with some discomfort,"First, I should say thatnot all the teams that I can think of, that were at a com-parable stage, ever did complete their task. About 40% ofthem eventually gave up. Ofthe remaining, I cannot thinkof any that completed their task in less than seven years,nor of any that took more than ten." He was then asked ifhe had reason to believe that the present team was moreskilled in curriculum development than the earlier oneshad been. "No," he replied,"I cannot think of any relevantfactor that distinguishes us favorably from the teams Ihave been thinking about. Indeed, my impression is thatwe are slightly below average in terms of resources andpotential."The wise decision at this point would probablyhave been for the team to disband. Instead, the members

ignored the pessimistic information and proceeded withthe project. They finally completed the initiative eightyears later, and their efforts went largely for naught-theresulting curriculum was rarely used.

In this example, the curriculum expert made two fore-casts for the same problem and arrived at very differentanswers. We call these two distinct modes of forecastingthe inside view and the outside view. The inside view is theone that the expert and all the other team members spon-taneously adopted. They made forecasts by focusingtightly on the case at hand-considering its objective, theresources they brought to it, and the obstacles to its com-pletion; constructing in their minds scenarios of theircoming progress; and extrapolating current trends intothe future. Not surprisingly, the resulting forecasts, even themost conservative ones, were exceedingly optimistic.

The outside view, also known as reference-class fore-casting, is the one that the curriculum expert was encour-aged to adopt. It completely ignored the details of theproject at hand, and it involved no attempt at forecastingthe events that would infiuence the project's futurecourse. Instead, it examined the experiences of a class ofsimilar projects, laid out a rough distribution of outcomesfor this reference class, and then positioned the currentproject in that distribution. The resulting forecast, as itturned out, was much more accurate.

The contrast between inside and outside views hasbeen confirmed in systematic research. Recent studieshave shown that when people are asked simple questionsrequiring them to take an outside view, their forecastsbecome significantly more objective and reliable. For ex-ample, a group of students enrolling at a college wereasked to rate their future academic performance relativeto their peers in their major. On average, these studentsexpected to perform better than 84% of their peers, whichis logically impossible. Another group of incoming stu-dents from the same major were asked about their en-trance scores and their peers' scores before being askedabout their expected performance. This simple detourinto pertinent outside-view information, which bothgroups of subjects were aware of, reduced the secondgroup's average expected performance ratings by 20%.That's still overconfident, but it's much more realistic thanthe forecast made by the first group.

Most individuals and organizations are inclined toadopt the inside view in planning major initiatives. It'snot only the traditional approach; it's also the intuitiveone. The natural way to think about a complex project isto focus on the project itself-to bring to bear all oneknows about it, paying special attention to its unique orunusual features. The thought of going out and gatheringstatistics about related cases seldom enters a planner'smind. The curriculum expert, for example, did not takethe outside view until prompted - even though he alreadyhad all the information he needed. Even when companies

JULY 2003 61


How to Takethe Outside View

Making a forecast using the outside

view requires planners to identify

a referenceclassofanalogous past initia-

tives, determine the distribution of out-

comes for those initiatives, and place the

project at hand at an appropriate point

along that distribution. This effort is best

organized into five steps:'

1. Select a reference class. Identifying

the right reference class involves both art

and science. You usually have to weigh

similarities and differences on many vari-

ables and determine which are the most

meaningful in judging how your own ini-

tiative will play out. Sometimes that's

easy. If you're a studio executive trying to

forecast sales of a new film, you'll formu-

late a reference class based on recent

films in the same genre, starring similar

actors, with comparable budgets, and so

on. In other cases, it's much trickier. If

you're a manager at a chemical company

that is considering building an olefin

plant incorporating a new processing

technology, you may instinctively think

that your reference class would include

olefin plants now in operation. But you

may actually get better results by looking

at other chemical plants built with new

processing technologies. The plant's out-

come, in other words, may be more influ-

enced by the newness of its technology

than by what it produces. In forecasting

an outcome in a competitive situation,

such as the market share for a new ven-

ture, you need to consider Industrial struc-

ture and market factors in designing a ref-

erence class. The key is to choose a class

that is broad enough to be statistically

meaningful but narrow enough to be truly

comparable to the project at hand.

2.Assess the distribution of outcomes.Once the reference class Is chosen, you

have to document the outcomes ofthe

prior projects and arrange them as a dis-

tribution, showing the extremes, the me-

dian, and any clusters. Sometimes you

won't be able to precisely document the

outcomes of every member ofthe class.

Butyou can still arrive at a rough distri-

bution by calculating the average out-

come as well as a measure of variability.

In the film example, for instance, you may

find that the reference-class movies sold

$40 million worth of tickets on average,

but that 10% sold less than $2 million

worth of tickets and 5% sold more than

$120 million worth.

3. Make an intuitive prediction of your

project's position in the distribution.Based on your own understanding ofthe

project at hand and how it compares with

the projects in the reference class, predict

where it would fail along the distribution.

Because your intuitive estimate will likely

be biased, the final two steps are intended

to adjust the estimate in order to arrive

at a more accurate forecast.

4. Assess the reliability of yourprediction. Some events are easier to

foresee than others. A meteorologist's

forecast of temperatures two days from

now, for example, will be more reliable

than a sportscaster's prediction ofthe

score of next year's Super Bowl. This step

is intended to gauge the reliabili^ ofthe

forecast you made in Step 3. The goal is

to estimate the correlation between the

forecast and the actual outcome, ex-

pressed as a coefficient between 0 and i,

where 0 indicates no correlation and l

indicates complete correlation. In the best

case, information will be available on how

well your past predictions matched the

actual outcomes. You can then estimate

the correlation based on historical prece-

dent. In the absence of such information,

assessmentsof predictability become

more subjective. You may,for instance,

be able to arrive at an estimate of pre-

dictability based on how the situation at

hand compares with other forecasting

situations. To return to the movie exam-

ple, say that you are fairly confident that

your ability to predict the sales of films

exceeds the ability of sportscasters to pre-

dict point spreads in football games but is

not as good as the ability of weather fore-

casters to predict temperatures two days

out. Through a diligent statistical analy-

sis, you could construct a rough scale of

predictability based on computed correla-

tions between predictions and outcomes

for football scores and temperatures. You

can then estimate where your ability to

predict film scores lies on this scale. When

the calculations are complex, it may help

to bring in a skilled statistician.

5. Correct the intuitive estimate.Due to bias, the intuitive estimate made

in Step 3 will likely be optimistic-devi-

ating too far from the average outcome

ofthe reference class. In this final step,

you adjust the estimate toward the aver-

age based on your analysis of predictabil-

ity in Step 4. The less reliable the predic-

tion, the more the estimate needs to be

regressed toward the mean. Suppose that

your intuitive predirtion ofa film's sales is

$95 million and that, on average,films in

the reference class do $40 million worth

of business. Suppose further that you have

estimated the correlation coefficient to be

0.6.The regressed estimate of ticket sales

would be:

$95 M + [0.6 C$40M-$95M)] = $62 M

As you see, the adjustment for optimism

will often be substantial, particularly in

highly uncertain situations where predic-

tions are unreliable.

^. This discussion builds on "tntuilive Prediaions: Biasesand Corrective Procedures," a 1979 article by DanielKahneman and Amos Tversky that appeared in TIMS Stud-ies in Management Science, volume la (Elsevier/NorthHolland).

HARVARD BUSINESS REVIEW


bring in independent consultants to assist in forecasting,they often remain stuck in the inside view. If the consul-tants provide comparative data on other companies orprojects, they can spur useful outside-view thinking. Butif they concentrate on the project itself, their analysiswill also tend to be distorted by cognitive biases.

While understandable, managers' preference for theinside view over the outside view is unfortunate. Whenboth forecasting methods are applied with equal intelli-gence and skill, the outside \iew is much more likely toyield a realistic estimate. That's because it bypasses cog-nitive and organizational biases. In the outside view, man-agers aren't required to weave scenarios, imagine events,or gauge their own levels of ability and control-so theycan't get all those things wrong. And it doesn't matter ifmanagers aren't good at assessing competitors' abilitiesand actions; the impact of those abilities and actions isalready reflected in the outcomes of the earlier projectswithin the reference class. It's true that the outside view,being based on historical precedent, may fail to predictextreme outcomes-those that lie out-side all historical precedents. But formost projects, the outside view will pro-duce superior results.

The outside view's advantage is mostpronounced for initiatives that compa-nies have never attempted before-likebuilding a plant with a new manufac-turing technology or entering an en-tirely new market. It is in the planningof such de novo efforts that the biasestoward optimism are likely to be great.Ironically, however, such cases are pre-cisely where the organizational and personal pressures toapply the inside view are most intense. Managers feel thatif they don't fully account for the intricacies ofthe pro-posed project, they would be derelict in their duties. In-deed, the preference for the iiiside view over the outsideview can feel almost like a moral imperative. The insideview is embraced as a serious attempt to come to gripswith the complexities of a unique challenge, while theoutside view is rejected as relying on a crude analogy tosuperficially similar instances. Yet the fact remains; Theoutside view is more likely to produce accurate forecastsand much less likely to deliver highly unrealistic ones.

Of course, choosing the right class of analogous casesbecomes more difficult when executives are forecastinginitiatives for which precedents are not easily found. It'snot like in the curriculum example, where many similarefforts had already been undertaken. Imagine that plan-ners have to forecast the results of an investment in anew and unfamiliar technology. Should they look at theircompany's earlier investments in new technologies? Orshould they look at how other companies carried outprojects involving similar technologies? Neither is perfect.

The outside viewis more iikeiy toproduce accurateforecasts andmuch less likely

unrealistic ones.

but each will provide useful insights-su the plannersshould analyze both sets of analogous cases. We provide afuller explanation of how to identify and analyze a refer-ence class in the sidebar "How to Take the Outside View"

Putting Optimism in Its PlaceWe are not suggesting that optimism is bad, or that man-agers should try to root it out of themselves or their or-ganizations. Optimism generates much more enthusiasmthan it does realism (not to mention pessimism), and itenables people to be resilient when confronting difficultsituations or challenging goals. Companies have to pro-mote optimism to keep employees motivated and fo-cused. At the same time, though, they have to generaterealistic forecasts, especially when large sums of moneyare at stake. There needs to be a balance betweeji opti-mism and realism-between goals and forecasts. Aggres-sive goals can motivate the troops and improve thechances of success, but outside-view forecasts should be

used to decide whether or not to makea commitment in the first place.

The ideal is to draw a clear distinc-tion between those functions and posi-tions that involve or support decisionmaking and those that promote or guideaction. The former should be imbuedwith a realistic outlook, while the latterwill often benefit from a sense of opti-mism. An optimistic CFO, for example,could mean disaster for a company, justas a lack of optimism would underminethe visionary qualities essential for su-

perior R&D and the esprit de corps central to a successfulsales force. Indeed, those charged with implementing aplan should probably not even see the outside-view fore-casts, which might reduce their incentive to perform attheir best.

Of course, clean distinctions between decision makingand action break down at the top. CEOs, unit managers,and project champions need to be optimistic and realisticat the same time. If you happen to be in one of these po-sitions, you should make sure that you and your plannersadopt an outside view in deciding where to invest amongcompeting initiatives. More objective forecasts will helpyou choose your goals wisely and your means prudently.Once an organization is committed to a course of action,however, constantly revising and reviewing the odds ofsuccess is unlikely to be good for its morale or perfor-mance. Indeed, a healthy dose of optimism will give youand your subordinates an advantage in tackling the chal-lenges that are sure to lie ahead. 9

Reprint R0307D; HBR OnPoint 4279To order, see page 119.

JULY 2003 63

Copyright 2003 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.

delusions of success

Documents