BEST PRACTICE

Why Good ProjectsFail Anyway

by Nadim F. Matta and Ronald N.Ashkenas

When a promising project doesn't deliver, chances

are the problem wasn't the idea but how it was

carried out Here's a way to design projects that

guards against unnecessary failure.

Bic PROJECTS FAIL at an astonish-ing rate. Whether major tech-nology installations, postmerger

integrations, or new growth strategies,these efforts consume tremendous re-sources over months or even years. Yetas study after study has shown, theyfrequently deliver disappointing re-turns - hy some estimates, in fact, wellover half the time. And the toll they takeis not just financial. These failures de-moralize employees who have laboreddiligently to complete their share of thework. One middle manager at a toppharmaceutical company told us, "I'vebeen on dozens of task teams in my ca-reer, and I've never actually seen onethat produced a result."

The problem is, the traditional ap-proach to project management shiftsthe project teams' focus away from theend result toward developing recom-mendations, new technologies, and par-tial solutions. The intent, of course, is topiece these together into a blueprintthat will achieve the ultimate goal, but

when a project involves many peopleworking over an extended period oftime, it's very hard for managers plan-ning it to predict all the activities andwork streams that will be needed. Un-less the end product is very well under-stood, as it is in highly technical engi-neering projects such as building anairplane, it's almost inevitable that somethings will be left off the plan. And evenif all the right activities have been an-ticipated, they may turn out to be diffi-cult, or even impossible, to knit togetheronce they're completed.

Managers use project plans, time-lines, and budgets to reduce what wecall "execution risk"-the risk that des-ignated activities won't be carried outproperly - but they inevitably neglectthese two other critical risks-the "whitespace risk"that some required activitieswon't be identified in advance, leavinggaps in the project plan, and the "inte-gration risk"thatthe disparate activitieswon't come together at the end. Soproject teams can execute their tasks

SEPTEMBER 2003109

BEST PRACTICE • Why Good Projects Fail Anyway

flawlessly, on time and under budget,and yet the overall project may still failto deliver the intended results.

We've worked with hundreds ofteams over the past 20 years, and we'vefound that by designing complex proj-ects differently, managers can reducethe likelihood that critical activities willbe left off the plan and increase theodds that all the pieces can be properlyintegrated at the end. The key is to injectinto the overall plan a series of mini-projects - what we call rapid-results ini-tiatives - each staffed with a team re-sponsible for a version of the hoped-foroverall result in miniature and each de-signed to deliver its result quickly.

Let's see what difference that wouldmake. Say, for example, your goal is todouble sales revenue over two years byimplementing a customer relationshipmanagement (CRM) system for yoursales force. Using a traditional projectmanagement approach, you might haveone team research and install softwarepackages, another analyze the differentways that the company interacts withcustomers (e-mail, telephone, and inperson, for example), another developtraining programs, and so forth. Manymonths later, however, when you startto roll out the program, you might dis-cover that the salespeople aren't sold onthe benefits. So even though they mayknow how to enter the requisite datainto the system, they refuse. This veryproblem has, in fact, derailed manyCRM programs at major organizations.

But consider the way the processmight unfold if the project includedsome rapid-results initiatives. A singleteam might take responsibility for help-ing a small number of users - say, onesales group in one region - increasetheir revenues by 25% within fourmonths. Team members would proba-bly draw on all the activities described

above, but to succeed at their goal, themicrocosm of the overall goal, theywould be forced to find out what, if any-thing, is missing from their plans as theygo forward. Along the way, they would,for example, discover the salespeople'sresistance, and they would be compelledto educate the sales staff about the sys-tem's benefits. The team may also dis-cover that it needs to tackle other issues,such as how to divvy up commissionson sales resulting from cross-selling orjoint-selling efforts.

When they've ironed out all the kinkson a small scale, their work would thenbecome a model for the next teams,which would either engage in furtherrapid-results initiatives or roll the sys-tem out to the whole organization -but now with a higher leve! of con-fidence that the project will have theintended impact on sales revenue. Thecompany would see an early paybackon its investment and gain new insightsfrom the team's work, and the teamwould have the satisfaction of deliveringreal value. ..-•*"

In the pages that follow, we'll take aclose look at rapid-results initiatives,using casd' studies to show how theseprojects are selected and designed andhow they are managed in conjunctionwith more traditional project activities.

How Rapid-ResultsTeams WorkLet's look at an extremely complex proj-ect, a World Bank initiative begun inJune 2000 that aims to improve the pro-ductivity of 120,000 small-scale farmersin Nicaragua by 30% in 16 years. A proj-ect of this magnitude entails many teamsworking over a long period of time, andit crosses functional and organizationalboundaries.

They started as they had always done:A team of World Bank experts and their

Nadim R Matta is a senior partner of Robert H. Schaffer & Associates, a managementconsulting firm based in Stamford, Connecticut. Prior to joining RHS&A, be workedfor USAID and tben beaded tbefood distribution program for Save the Children dur-ing the civil war in Lebanon. Ronald N. Ashkenas is a managing partner ofRHS&A.He has written tbree previous articles for HBR, most recently "Integration Managers:Special Leaders for Special Times" (witb Suzanne C. Francis, November-December2000.) Tbe authors can be reached at info@rbsa.com.

clients in the country (in this case. Min-istry of Agriculture officials) spent manymonths in preparation - conductingsurveys, analyzing data, talking to peo-ple with comparable experiences inother countries, and so on. Based ontheir findings, these project strategists,designers, and planners made an edu-cated guess about the major streams of

Managers expect they

can plan for all the

variables in a complex

project in advance, but

they can't Nobody is

that smart or has that

clear a crystal ball.

work that would be required to reachthe goal. These work streams includedreorganizing government institutionsthat give technical advice to farmers,encouraging the creation of a private-sector market in agricultural supportservices (such as helping farmers adoptnew farming technologies and use im-proved seeds), strengthening the Na-tional Institute for Agricultural Tech-nology (INTA), and establishing aninformation management system thatwould help agricultural R&D institu-tions direct their efforts to the most pro-ductive areas of research. The result ofall this preparation was a multiyearproject plan, a document laying out thework streams in detail.

But if the World Bank had kept pro-ceeding in the traditional way on a proj-ect of this magnitude, it would havebeen years before managers found outif something had been leff off the planor if the various work streams couldbe integrated - and thus if the projectwould ultimately achieve its goals. Bythat time, millions of dollars would havebeen invested and much time poten-tiaUy wasted. What"s more, even if every-thing worked according to plan, theproject's beneficiaries would have beenwaiting for years before seeing any pay-off ffom the effort. As it happened, the

110 HARVARD BUSINESS REVIEW

Why Good Projects Fail Anyway • BEST PRACTICE

project activities proceeded on sched-ule, but a new minister of agriculturecame on board two years in and arguedthat he needed to see results soonerthan the plan allowed. His complaintresonated with Norman Piccioni, theWorld Bank team leader, who was alsogetting impatient with the project'space. As he said at the time,"Apart fromthe minister, the farmers, and me, I'mnot sure anyone working on this projectis losing sleep over whether fanner pro-ductivity will be improved or not"

Over the next few months, we workedwith Piccioni to help him and his clientsadd rapid-results initiatives to the im-plementation process. They launchedfive teams, which included not only rep-resentatives from the existing workstreams but also the beneficiaries ofthe project, the farmers themselves. Theteams differed from traditional imple-mentation teams in three fundamentalways. Rather than being partial, hori-zontal, and long term, they were resultsoriented, vertical, and fast. A look ateach attribute in turn shows why theywere more effective.

Results Oriented. As the name sug-gests, a rapid-results initiative is inten-tionally commissioned to produce ameasurable result, rather than recom-mendations, analyses, or partial solu-tions. And even though the goal is ona smaller scale than the overall objec-tive, it is nonetheless challenging. InNicaragua, one team's goal was to in-crease Grade A milk production in theLeon municipality from 600 to 1,600gallons per day in 120 days in 60 smalland medium-size producers. Anotherwas to increase pig weight on 30 farmsby 30% in too days using enhanced comseed. A third was to secure commit-ments from private-sector experts toprovide technical advice and agricul-tural support to 150 small-scale farmersin the El Sauce (the dry farming region)within 100 days.

This results orientation is importantfor three reasons. First, it allows projectplanners to test whether the activitiesin the overall plan will add up to theintended result and to alter the plansif need be. Second, it produces real ben-

efits in the short term. Increasing pigweight in 30 farms by 30% in just overthree months is usefril to those 30 farm-ers no matter what else happens in theproject. And finally, being able to de-liver results is more rewarding and en-ergizing for teams than plodding alongthrough partial solutions.

The focus on results also distinguishesrapid-resuits initiatives from pilot proj-ects, which are used in traditionally

an organization (or even, as in Nicara-gua, different organizations), and thevertical slice brings these people to-gether. This vertical orientation is keyto reducing white space and integrationrisks in the overall effort: Only by un-covering and properly integrating anyactivities falling in the white space be-tween the horizontal project streamswill the team be able to deliver its mini-result (For a look at the horizontal and

managed initiatives only to reduce exe-cution risk. Pilots typically are designedto test a preconceived solution, or means,such as a CRM system, and to work outimplementation details before rollout.Rapid-results initiatives, by contrast, areaimed squarely at reducing white spaceand integration risk.

Vertical. Project plans typically un-fold as a series of activities representedon a timeline by horizontal bars. In thiscontext, rapid-results initiatives are ver-tical. They encompass a slice of severalhorizontal activities, implemented intandem in a very short time frame. Byusing the term "vertical," we also sug-gest a cross-functional effort, since dif-ferent horizontal work streams usuallyinclude people from different parts of

vertical work streams in the Nicaraguaproject, see the exhibit "The WorldBank's Project Plan.")

The team working on securing com-mitments between farmers and tech-nical experts in the dry farming region,for example, had to knit together abroad set of activities. The expertsneeded to be trained to deliver par-ticular services that the farmers weredemanding because they had heardabout new ways to increase their pro-ductivity through the information man-agement system. That, in turn, was be-ing fed information coming out of I NTA'sR&D efforts, which were directed to-ward addressing specific problems thefanners had articulated. So team mem-bers had to draw on a number of the

SEPTEMBER 2003 111

BEST PRACTICE • Why Good Projects Fait Anyway

broad horizontal activities laid out inthe overall project plan and integratethem into their vertical effort. As theydid so, they discovered that they had toadd activities missing from the originalhorizontal work streams. Despite theteam members' heroic efforts to inte-grate the ongoing activities, for instance,80 days into their 100-day initiative, theyhad secured only half the commitmentsthey were aiming for. Undeterred andspurred on by the desire to accomplishtheir goal, team members drove throughthe towns of the region announcingwith loudspeakers the availability andbenefits of the technical services. Overthe following 20 days, the gap to thegoal was closed. To close the white spacein the project plan, "marketing of tech-nical services" was added as anotherhorizontal stream.

Fast How fast is fast? Rapid-resultsprojects generally last no longer than100 days. But they are by no meansquick fixes, which imply shoddy orshort-term solutions. And while theydeliver quick wins, the more importantvalue of these initiatives is that theychange the way teams approach theirwork. The short time frame fosters asense of personal challenge, ensuringthat team members feel a sense of ur-gency right from the start that leavesno time to squander on big studies orinterorganizational bickering. In tradi-tional horizontal work streams, the gapbetween current status and the goalstarts out far wider, and a feeling of ur-gency does not build up until a shorttime before the day of reckoning. Yet itis precisely at that point that committedteams kick into a high-creativity modeand begin to experiment with new ideasto get results. That kick comes rightaway in rapid-results initiatives.

A Shift in AccountabilityIn most complex projects, the execu-tives shaping and assigning major workstreams assume the vast majority of theresponsibility for the project's success.They delegate execution risk to projectteams, which are responsible for stay-ing on time and on budget, but they in-advertently leave themselves carrying

the full burden of white space and inte-gration risk. In World Bank projects, asin most complex and strategically criti-cal efforts, these risks can be huge.

When executives assign a team re-sponsibility for a result, however, theteam is free - indeed, compelled - tofind out what activities will be neededto produce the result and how thoseactivities will fit together. This approachputs white space and integration riskonto the shoulders of the people doingthe work. That's appropriate because, asthey work, they can discover on the spotwhat's working and what's not. And inthe end, they are rewarded not for per-forming a series of tasks but for deliver-ing real value. Their success is correlatedwith benefits to the organization, whichwill come not only from Implementingknown activities but also from identify-ing and integrating new activities.

The mi!l< productivity team in Nica-ragua, for example, found out early onthat the quantity of milk productionwas not the issue. The real problem wasquality: Distributors were being forcedto dump almost half the milk they hadbought due to contamination, spoilage,and other problems. So the challengewas to produce milk acceptable to largedistributors and manufacturers thatcomplied with international qualitystandards. Based on this understanding,the team leader invited a representativeof Parmalat, the biggest private com-pany in Nicaragua's dairy sector, to jointhe team. Collaborating with this cus-tomer allowed the team to understandParmalat's quality standards and thusintroduce proper hygiene practices tothe milk producers in Leon. The collab-oration also identified the need for sim-ple equipment such as a centrifuge thatcould test the quality of batches quickly.

The quality of milk improved steadilyin the initial stage of the effort. But thenthe team discovered that its goal oftripling sales was in danger due to a lo-gistics problem: There wasn't adequatestorage available for the additionalGrade A milk now being produced.Rather than invest in refrigeration fa-cilities, the Parmalat team member(now assured of the quality of the milk)

suggested that the company conductcollection runs in the area daily ratherthan twice weekly.

At the end of 120 days, the milk pro-ductivity team (renamed the "clean-milldng"team) and the other four teamsnot only achieved their goals but alsogenerated a new appreciation for thediscovery process. As team leader Pic-cioni observed at a follow-up workshop:"I now realize how much of the overallsuccess of the effort depends on peoplediscovering for themselves what goalsto set and what to do to achieve them."

What's more, the work is more re-warding for the people involved. It mayseem paradoxical, but virtually all theteams we've encountered prefer to workon projects that have results-orientedgoals, even though they involve somerisk and require some discovery, ratherthan implement clearly predefined tasks.

The LeadershipBalancing ActDespite the obvious benefits of rapid-results initiatives, few companies shoulduse them to replace the horizontal ac-tivities altogether. Because of theireconomies of scale, horizontal activitiesare a cost-efficient way to work. And soit is the job of the leadership team tobalance rapid-results initiatives withlonger-term horizontal activities, helpspread insights from team to team, andblend everything into an overall imple-mentation strategy.

In Nicaragua, the vertical teams drewmembers from the horizontal teams,but these people continued to work onthe horizontal streams as well, and eachteam benefited from the work of theothers. So, for example, when the milkproductivity team discovered the needto educate farmers in clean-milkingpractices, the horizontal training teamknew to adjust the design of its overalltraining programs accordingly.

The adhesive-material and office-product company Avery Dennison tooka similar approach, creating a portfolioof rapid-results initiatives and horizon-tal work streams as the basis for its over-all growth acceleration strategy. Just overa year ago, the company was engaged in

112 HARVARD BUSINESS REVIEW

Why Good Projects Fail Anyway • BEST PRACTICE

The World Bank'sProject PlanA project plan typically represents the planned activities

as horizontal bars plotted overtime. But in most cases,

it's very difficult to accurately assess all the activities that

will be required to complete a complicated long-term

project. We don't know what will fall into the white space

between the bars. It's also difficult to know whether these

activities can be integrated seamlessly at the end; the

teams working in isolation may develop solutions that

won't fit together. Rapid-results initiatives cut across hori-

zontal activities, focusing on a miniversion of the overall

result rather than on a set of activities.

Here is a simplified version of the Nicaragua project

described in this article. Each vertical team (depicted as

a group by the vertical bar) includes representatives from

every horizontal team, which makes the two types of initia-

tives mutually reinforcing. So, for example, the horizontal

work stream labeled "Set up private-sector market in agri-

cultural support services" includes activities like develop-

ing a system ofcoupons to subsidize farmers'purchases.

The vertical team establishing service contracts between

technical experts and farmers drew on this work, providing

the farmers with coupons they could use to buy the techni-

cal services. This, in turn, drove competition in the private

sector, calling on the work that the people on the horizontal

training teams were doing -which led to better services.

Overall Project Objective:Improve productivity of 120,000farmers by 30% in 16 years

.Reorganize government agriculturaltechnical-service institutions

Set up private-sector marketin agricultural support services

Long-term ' Strengthen National Institutework streams for AgriculturalTechnology

Implement training programs foragricultural technical-service provider

Establish agricultural informationmanagement system

for p t t i ^ - ^ - ^ a n alternativepurchase agreement^ ' '

2000 Rapid-results initiatives(drawing on the work of all long-term work streams)

^2016

SEPTEMBER 2003113

BEST PRACTICE • Why Cood Projects Fail Anyway

various horizontal activities like newtechnology investments and marketstudies. The company was growing, butCEO Phil Neal and his leadership teamwere not satisfied with the pace. Al-though growth was a major corporategoal, the company had increased its rev-enues by only 8% in two years.

In August 2002, Neal and presidentDean Scarborough tested the verticalapproach in three North American di-visions, launching 15 rapid-results teamsin a matter of weeks. One was chargedwith securing one new order for an en-hanced product, refined in collaborationwith one large customer, within lOOdays. Another focused on signing upthree retail chains so it could use thatexperience to develop a methodology

Rapid-results initiatives

challenge senior leaders

to cede control.

for moving into new distribution chan-nels. A third aimed to book several hun-dred thousand dollars in sales in 100days by providing - through a collabo-ration with three other suppliers - allthe parts needed by a major customer.By December, it had become clear thatthe vertical growth initiatives were pro-ducing results, and the managementteam decided to extend the processthroughout the company, supported byan extensive employee communicationcampaign. The horizontal activities con-tinued, but at the same time dozens ofteams, involving hundreds of people,started working on rapid-results initia-tives. By the end of the first quarter of2003, these teams yielded more than$8 million in new sales, and the com-pany was forecasting that the initiativeswould realize approximately $50 mil-lion In sales by the end of the year.

The Diversified Products businessof Zurich North America, a division ofZurich Financial Services, has taken asimilarly strategic approach. CEO RobFishman and chief underwriting offi-cer Gary Kaplan commissioned andlaunched dozens of rapid-results initia-

tives between April 1999 and Decem-ber 2002. Their overall long-term ob-jectives were to improve their financialperformance and strengthen relation-ships with core clients. And so they com-bined vertical teams focused on suchgoals as increasing payments from asmall number of clients for value-addedservices with horizontal activities tar-geting staff training, internal processes,and the technology infi-astructure. Theresults were dramatic: In less than fouryears, loss ratios in the property side ofthe business dropped by 90%, the ex-pense ratio was cut in half, and fees forvalue-added services increased tenfold.

Now, when you're managing a port-folio of vertical initiatives and horizon-tal activities, one of the challenges be-comes choosing where to focus theverticals. We generally advise companyexecutives to identify aspects of theeffort that they're fairly sure will fail ifthey are not closely coordinated withone another. We also engage the lead-ership team in a discussion aimed atidentifying other areas of potentialuncertainty or risk. Based on those dis-cussions, we ask executives to think ofprojects that could replicate their longer-term goals on a small scale in a shorttime and provide the maximum oppor-tunity for learning and discovery.

For instance, at Johnson & Johnson'spharmaceutical R&D group, ThomasKirsch, the head of global quality assur-ance, needed to integrate the QA func-tions for two traditionally autonomousclinical R&D units whose people werelocated around the world. Full inte-gration was a major undertaking thatwould unfold over many years, so inaddition to launching an extensive se-ries of horizontal activities like devel-oping training standards and devisinga system for standardizing currently dis-parate automated reports, Kirsch alsoassigned rapid-results teams to quicklyput in place several standard operatingprocedures (SOPs) that cut across thehorizontal work streams. The rapid-results teams were focused on the areashe perceived would put the company inthe greatest danger of failing to complywith U.S. and European regulations and

also on areas where he saw opportuni-ties to generate knowledge that couldbe applied companywide. There's no sci-ence to this approach; it's an iterativeprocess of successive approximation, nota cut-and-dried analytical exercise.

In fact, there are really no "wrong"choices when it comes to decidingwhich rapid-results initiatives to add tothe portfolio. In the context of a large-scale, multiyeai, high-stakes effort, eachlOQ-day initiative focused on a targetedresult is a relatively low-risk investment.Even if it does not fully realize its goal,the rapid-results initiative will producevaluable lessons and help further illu-minate the path to the larger objective.And it will suggest other, and perhapsbetter-focused, targets for rapid results.

A Call for HumilityRapid-results Initiatives give some newresponsibilities to frontline team mem-bers while challenging senior leaders tocede control and rethink the way theysee themselves. Zurich North America'sGary Kaplan found that the process ledhim to reflect on his role."l had to learnto let go: Establishing challenging goalsand giving others the space to figure outwhat it takes to achieve these...did notcome naturally to me."

Attempting to achieve complex goalsin fast-moving and unpredictable envi-ronments is humbling. Few leaders andfew organizations have figured out howto do it consistently. We believe that astarting point for greater success is shed-ding the blueprint model that has im-plicitly driven executive behavior in themanagement of major efforts. Manag-ers expect they will be able to identify,plan for, and influence all the variablesand players in advance, but they can't.Nobody is that smart or has that cleara crystal ball. They can, however, createan ongoing process of learning and dis-covery, challenging the people close tothe action to produce results - and un-leashing the organization's collectiveknowledge and creativity in pursuit ofdiscovery and achievement. ^

Reprint R0309H; HBR OnPoint 4872To order, see page 135-

114 HARVARD BUSINESS REVIEW