Stage-Gate Systems: A NewTool for Managing NewProductsRobert G. Cooper

New products conhelp your oomponymuch more quicklyand efficiently with abit of planning beforedevelopment starts.

A merica is in aproduct war,and the man-agement of innova-tion is the strategicweapon. The battlestake place on manyfields: from electronicchips to tractors, cam-eras to machine tools,scientific instrumentsto automobiles.

Our ability toget better at the inno-vation processto

drive new products from idea to market fasterand with fewer mistakesis the key to winningthis war. Sadly, many firms miss the mark: Onlyone product development project in four be-comes a winner, and almost 50 percent of theresources American firms devote to innovationare spent on products that are commercial fail-ures.

Stage-gate systems form one solution to whatails many firms' new product programs. Facingincreased pressure to reduce the cycle time yetimprove their new product "hit rate," corpora-tions are increasingly looking to stage-gate mod-els as effective tools to manage, direct, and con-trol their product-innovation efforts.

What are stage-gate systems? A stage-gatesystem is both a conceptual and an operationalmodel for moving a new product from idea tolaunch. It is a blueprint for managing the new-product process to improve effectiveness andefficiency. Although conceptually quite simple, aswe shall see later, the intricacies, design, andoperationalization of stage-gate approaches areconsiderably more complex.

THE NEED FOR BETTER NEW PRODUCTMANAGEMENT

Facing increased competition from homeand abroad, maturing markets, and theheightened pace of technological change,corporations look to new products and new busi-nesses for sustained growth and competitive ad-vantage.

The desire to develop and launch new prod-ucts is obvious. The manager's bookshelf is re-plete with new books on managing inniwation,launching new products, and managing technol-ogy. A .study by the Conference Board (Hopkins1980) revealed that, by an eight-to-one ratio,CEOs believed that their firms would be muchmore dependent on new products in the yearsahead. A Coopers & Lybrand survey (1985) re-ported that most companies are counting heavilyon new product development for growth andprofitability.

There is good reason f(jr this heightenedinterest in product innovation. An annual Fortunesurvey rates top American corporations on anumber of criteria, including "value as a long-term investment." Using data supplied hyFortune.we studied various predictors of investmentvalue. The results were provocative: The singlestrongest predictor of inve.stment value is "degreeof innovativeness of the company." A typicalindustry relationshiphow innovativeness im-pacts on investment valueis shown for thechemical industiy in Figure 1.

The need for effective product innovation isthere; are the results? Products continue to fail atan alarmingly high rate. In one study (Hopkins1980), 63 percent of managers felt that their newproduct success rate was "disappointing" or "un-acceptably low." The Coopers Sc Lybrand survey

44 Busines.s Horizoas / May-Junt" 199()

revealed that there is widespread disillu-sionment with many t)f the new productsthat firms develop (notably extensions andincremental improvements) and that firmshave had too much of a "tech push" ratherthan "market pull" orientation in their newproduct efforts. The inability of many firmsto use their internal resources effectively fornew product growth may be one reason forthe recent merger mania, itself almost anadmission of the failure of their internalgrowth programs.

THE SOLUTION

T he strategic solution is that manage-iiienL must get better at conceiving,developing, and launching newproducsnot just extensions and incre-mental improvements, but new productsthat givtr tlie firm a sustainable competitiveadvantage. Ulis translates into better man-agement of the innovation process. Stage-gate systems are seen as one answer, as thefollowing examples show.

With a rather dismal new products andmodels record for the last decade, GeneralMotors is now trying to beat the Japanese attheir own game. GM is currently imple-menting a Four Phase system for producttiesign and introductiona methodology thatpromises to drastically cut the idea-to-launch timeof a new car model, This copyrighted Four Phasesystem is nothing more than GM's version of astage-gate system,

3M traditionally has had an enviable newproduct track record. An innovative corporateculture and climate are often cited as 3M's secretweapons. But 3M has also had in place variousstage-gate systems for managing the innovationprocess. Thus creativity and discipline are blend-ed to yield a successful new product program.

One major packaged goods firm, alwaysnoted for its forward-thinking inanagement meth-ods, is currently facing tougher times in its newproducts efforts. A senior committee has tried tofind out why. One fact was that back in 1964. thecompany had implemented a gating system at atime when most people thought that productinnovation couldn't be managed. The systemworked welt for 15 years, but fell from favor inthe late 197S. One reccmmendation of the com-mittee: "Let's get back to basics and redesign andre-implement the stage-gate process."

Noithern Telecom, a telecommunicationsequipment manufacturer who has successfullypenetrated the international market in recentyears, implemented their four-,siage gating systemf't>r new products in the 1980s. The stage-gatemodel cost approximately $1 million to design

Figure 1How Innovativeness Affects Value as an Investment:Chemical Industry

6 -

c

Allied

HerculesFMC

Celanese

Innovativeness

Source: Fortune and Eraos & Morgan. Inc.

and implement, but the results have been impres-sive; shorter times to launch; fewer mistakes; le,ssrecycling and rework in the process; and a moresuccessful development effort.

A major electrical/electronics corporationrecently undertook a study of the innovationprocess within its roughly 50 operating divisions.The results were conclusive: Only a handful ofthe divisions had implemented stage-gate sys-tems, but those few were achieving a muchhigher level of new product performance thanthose divisions that lacked a formal game plan.

THE CONCEPT OF A STAGE-GATE SYSTEM

I ndividual companies may refer to their sys-tems by different names, and on paper theyappear to be unique to that company. Inpractice, however, there is a surprising parallel-ism between different stage-gate approaches.

Stage-gate sy.stems recognize that productinnovation is a process. And like other processes,innovation can be managed. Stage-gate systemssimply apply process-management methodologiesto this innovation process.

A good analogy is the production process tomanufacture a physical product. The way to im-prove the quality of output from the process, ofcourse, is to focus on the process itselfto re-move variances in the process, A process is sub-

Stage-Gaie Systems; A New Tool for Managing New Products

Figure 2An Overview of a Stage-Gate System

PreliminaryAssessment

DetailedInvestigation

(Business Case)Preparation

Development Testing &Validation

FullProduction& MarketLaunch

Stage1

Stage2

Idea InitialScreen

SecondScreen

Stage3

Deci.sionon

BusinessCase

Stage

Post-Development

Review

Stage

cializationBusinessAnalysis

Post-Implemen-

tationReview

divided into a number of stages or work stations.Between each work station or ,stage, there is aquality control checkpoint or gate, A set of deliv-erables is specified for each gate, as is a set ofquality criteria that the product must pass beforemoving to the next work station. The stages arewhere the work is done; the gates ensure that thequality is sufficient.

Stage-gate ,systems use similar metliods tomanage the innovation process. They divide theinnovation process into a predetermined set ofStages, themselves composed of a group of pre-scribed, related, and often parallel activities. Forexample, the "Validation" stage might entail a listof mandatory or optional activities such as in-house prototype tests, field tests with customers,pilot or trial production, and test marketing.

Usually stage-gate systems involve from fourto seven stages and gates, depending on thecompany or division, A typical system is shownin Figure 2, Each stage is usually more expen-sive than the preceding one. Concurrently, infor-mation becomes better and better, so risk is man-aged.

The entrance to each stage is a gate; thesegates control the process, much like quality con-trol checkpoints control the production process.Each gate is characterized by a set of deliverablesor inputs, a set of exit criteria, and an output.The inputs are the deliverables that the projectleader must bring to the gate. The criteria are theitems upon which the project will be judged, thehurdles that the project must pass at that gate tohave the gate opened to the next stage. The out-puts are the decisions at the gate, typically a Go/Kill/Hold/Recycle decision, and the approval ofan action plan for the next stage.

Each project leader is required tt) provide thespecified deliverables and meet the stated criteriaat a given gate. For example, at Gate 3 in Figure2, the inputs might include: results of the u,ser"needs and wants" market study; the competitiveanalysis; the detailed technical appraisal; and thefinancial as,sessment. The inputs and the criteriachange from gate to gate; Gate l's inputs andcriteria are quite different from Gate 4 s.

Gates are manned by senior managers whoact as "gatekeepers," This gatekeeping group istypically multidisciplinary and multifunctional,and its members are senior enough to have theauthority to approve the resources needed by theproject. Its role includes:

Review of the quality of the inputs or de-liverables;

Assessment of the quality of the projectfrom an economic and business standpoint, re-sulting in a Go/Kill/Hold/Recycle decision; and

Approval of the action plan for the nextstage (in the event of a Go decision) and alloca-tion of the necessary resources.

The project leader drives the project fromstage to stage, gate to gate. He or she is wellaware of what inputs are required to "pass" thenext gate and organizes the team to meet theinput requirements of the upcoming gate.

The implementation of stage-gate systemsrequires certain organizational changes withinsome firms. For example, a project team ap-proach to organizing new product projects is fun-damental to stage-gate approaches. No longercan projects be handed from department to de-partment within the firm; a team and leader mustcany the project in ail stages,

A second organizational change for some

46 Business Horizons / May-June 1990

finns is the involvement of senior management asgatekeepers. Successful produa innovation re-quires significant resources and demands thecommitment of top management. Gates mannedby senior people are not only essential to gate-ways systems; they build in top management in-volvement and commitment.

WHERE'S THE EVIDENCE?

A number of firms have adopted such afonnal model. But does such a processapproach, although intuitively appealing,have any impact on performance? One study(Booz, Allen & Hamilton 1982) found that it does:firms that adopted a formal new product processdid better, and those firms with the process inplace for the longest time fared the best.

A Quality Focus

lhe issue of a successful new product processone that is complete and is executed in a qualityfashionmust be a fundamental concern of man-agement. But a research study of 203 new prod-uct projects (Cooper and Kleinschmidt 1986)

suggests that quality processes are lacking inmost firms' new product programs. New productprocesses were found to be incomplete and tosuffer from sloppy or under-resourced execution.

In three-quarters of these 205 projects,there was no detailed market study or marketingresearch undertaken at all. Yet a lack of marketassessment has been consistently cited for yearsas the number-one reason for new product fail-ure. Fully 77 percent of these projects featuredno test market or trial sell; 34 percent omittedproduct testing with the customer; and 32 percentof these new product projects did not even havea formal product launch.

Thirteen commonly prescribed activities werestudied in these 203 projects, ranging from "initialscreening" to "prototype/sample tests" and "for-mal launch." Only four projects out of the 203featured all 13 activitiesthat is, could be consid-ered a complete or "textbook" approach. Theother 98 percent were abbreviated processes,(iften omitting crucial steps. Further, the studyfound that the completeness of the process wassignificantly related to project success: the moresteps or activities one left out, the higher thelikelihood of failure.

Figure 3Quality of Execution of ActivitiesSuccess versus Failure

- Initial Screening

Preliminary Market Assessment

- Preliminary Technical Assessment

- Detailed Market Study/Marketing Research

- Business/Financial Analysis

- Product Development

In-House Product Testing

Customer Test of Product

- Test Market/Trial Sell

" Trial Production

- Pre-Commercialization Business Analysis

Production Start-Up

- Market Launch

Rating

Stage-Gale Sysieins: A New Tool for Managing New Products 47

"The most pivotal activi-ties, those in which thedifferences betweensuccesses and faiiureswere the greatest, werethe early activities in thenew product process. "

A similar picture emerges when one looks atquality of execution of each of these 13 activities(see Figure 3)- Successful products were stronglylinked to quality of execution. Where activitieswere proficiently undertaken, there was a higherlikelihood of new product success. Those mostpoorly executed, and those cited in greatest need

of improvement, wereinitial screening, thedetailed market study,and preliminary mar-ket assessment. Tliemost pivotal activities,those in which thedifferences betweensuccesses and failureswere the greatest,were the early activi-ties in the new prod-uct process. Theseeds of success orfailure are sown inthe first few steps of

the process: the predevelopment or "homework"stages.

What this and other studies reveal is convinc-ing. First, most products fail because of errors ofomission and commission in the new productprocess: a lack of market assessment; productdefects; inadequate launch efforts; poor screeningand project evaluation, and so on. Second, cur-rent performance of product innovation is farfrom ideal: there are too many missing steps andshort-cuLs, and questionable quality of execution.Third, quality of execution separates winnersfrom losers; How well the activities are carriedout, and indeed whether they are carried out atall, is strongly correlated with project outcomes.And finally, the weakest and most pivotal activi-ties are the predevelopment and market-orientedones.

Stage-gate systems, although simple ccincep-tually, have a profound impact on the innovationprocess. Stage-gate models provide the qualityfocus that is often missing in firms' new productprograms. By building in quality control check-points in the form of gates, stage-gate systemsensure that project leaders and teams meet highstandards of execution. As the project leaderapproaches a gate, he or she knows what inputsare required and that these deliverables will becarefully scrutinized by the gatekeepers. Thepressure is very much on the project leader tobuild quality into his or her project.

Gates ensure that no critical activities havebeen omitted: an action plan is agreed upon ateach gate, and the deliverables for the next gateare cleady specified. The result is no critical er-rors of omission, no gaps in the process, and a"complete" process.

Attention and resources are devoted to activi-ties and stages that are often deficient in the in-novation process. Stage-gate systems typicallyemphasize a market orientation and marketinginputs, and they devote far more attention to thefront-end or "homework" stages that precede theproduct development phase.

A Stronger Market Orientation

A lack of a market orientation and inadequatemarket a.ssessment are consistently cited as majorreasons for new product failure, panicuiarly inindustrial-product and high-technology firms.Moreover, the market-oriented activities tend tobe the weakest in the new product process, yetthey are strongly linked to .success.

Many managements profess a market orienta-tion, but the evidence is otherwise in the case ofnew products. In one study of new product casehistories (Cooper and Kleinschmidt 1988). only16 percent of the total effort expended on newproduct projects went to market-oriented activi-ties. The breakdown for the average project inman-days spent was:

Technical and production activities, 78percent;

Market-oriented activities (includinglaunch). l6 percent;

Evaluative/financial activities, 6 percent.If launch activities are removed, the amount

devoted to efforts such as market as.sessment,detailed market studies, customer tests, and trialsell/test markets becomt-s pitifully small.

The provocative finding is that those firmsthat did proportionately more market-relatedactivities reaped the benefiLs. Overall, successfulnew products had considerably more time,money, and energy devoted to market-orientedactivities than did failures. In successful projects,three times as many man-days and twice as muchmoney were devtjted to preliminary market as-sessment than was the case for faiiures. Twice asmuch market research (measured in both man-days and dollars spent) was conducted in suc-cessful products as in failures. But in both cases,the amounts were still small. Successful productshad more than twice as much money spent oncustomer tests of the product as did failures. Andsix times as much money and twice as manyman-days were spent on the launch of successfulproducts as for failures.

None of these marketing costs was significantrelative to the total project costs, with the excep-tion of launch. So doubling the marketing effortdid not materially affect the total project costs butclearly yielded a marked impact on project out-comes. Stage-gate systems provide for a muchstronger market orientation in the new productprocess. Those market-related activities, so often

Business Horizons / May-June 1990

omitted or weakly handled in most new product[projects, are built intt) the process by design, notas an afterthought. The .stages of the processtypically include a number of market-relatedactivities, such as user needs and wants research;concept tests; competitive analysis; development(if a detailed marketing plan; product tests withcustomers; Irial sell; and fonnal launch. The proj-ect leader must ensure that these critical steps areexecuted: unless they are, his or her project doesnot pass the next gate.

Better Homework

We all learned in Fifth grade how distastefulhomework is. (Many of us haven't forgotten.) Buthomework is critical to a successful developmentprogram. Sadly, too many new product projectssuffer from a lack of homework. Only 7.1 percentof the total expenditures on the typical projectwas spent on homework (or predevelopment)activitiesactivities undertaken to qualify anddefine the project prior to a major developmentprogram. By contrast, development and producttesting received 39 percent of expenditures, andcommercialization 54 percent. In man-days spent,only 16.4 percent of the total project effort wentto predeveiopment efforts. Development andproduct testing received 61 percent and commer-cialization 22,5 percent. Of all the activities in thenew product proce.ss, predevelopment activitieswere the ones most weakly executed and ingreatest need of improvement.

The most important steps of the new productprocessthose that usLially separate winnersfrom ltjserslie in the stages that precede theproduct development pha.se. Figure 3, whichcaptures the quality of execution in typical proj-ects, reveals that how well these homeworkstages are undertaken is strongly linked to prod-uct success. Further, successful projects spendover twice as much money and 1.75 times asmany man-tlays on predevelopment steps asfailures. Finally, a Booz, Allen k Hamilton studyU9H2) found that Japanese firms and successfulU.S. firms devote considerably more time to thehomework stages before entering developmentthan does the average U.S. firm.

The predeveiopment activities are importantbecause they qualify and define the project. Theyanswer key questions such as:

Is the project an economically attractiveone? Will the product sell at sufficient volumes.iiid margins to justify the investment in develop-ment and commercialization?

Who exactly is the target customer? Andlu)w sht)uld the product be positioned?

What exactly should the product be tomake it a winner? What features and attributesshould be built into the product to give it a dif-

ferential product advantage? Can the product be developed, and at the

right costs? What is the likely technical solution?Stage-gate systems provide a focus on home-

work. The typical game plan has one or twostages devoted to predevelopment investigation,as shown in Figure 2. If this homework is notdone or is done poorly, the project fails to enterthe expensive development phase.

"More homework means longer developmenttimes" is a frequently voiced complaint. This is avalid concern, but experience has shown that thishomework pays for itself in reduced develop-ment time and improved success rates. First,without homework, there is a much higher likeli-hood of new product failure. So the choice isbetween a slightly longer project or increasedodds of failure.

Second, many projects are poorly definedwhen they enter the development phase. This isoften the result of weak predevelopment activi-ties: the target user is not well understood, userneeds and wants are vaguely defined, and re-quired product features and attributes are fuzzy.R&D people and design engineers are not mindreaders. With a poorly defined project, theywaste considerable time seeking definition, oftenrecycling back severaltimes to "get the productright." Better projectdefinition, the result ofsound homework, actu-ally speeds up the de-velopment process.

Third, rarely does aproduct concept remainthe same from begin-ning to end. The originalidea that triggered theproject is seldom thesame as what goes tomarket. Given this inevi-table product design evolution, the time to makethe majority of these design changes is not whenthe product is moving out of development andinto production. More homework up front en-courages changes to occur earlier in the processrather than later, when they are more costly. Theresult is a considerable savings in time andmoney at the back end of the project and a moreefficient new product process.

Parallel Processing

Parallel processing is an important feature ofstage-gate systems. It means that activities areparallel rather than .sequential. At each stage ofthe gateways sy.stem, many activities take placeconcurrently and involve different functions ofthe firm.

"More homework upfront encourageschanges to occurearlier in the processrather than later,when they are morecostly "

Stage-Gate Systems: A New Tool for Managing New Products 49

why parallel processing? New product man-agers face a dilemma: On one hand, they arebeing urged by corporate management to shortenthe elapsed time from idea to launch. On theother hand, the manager is urged to improve theeffectiveness of product development: cut downthe failure rate; do it right. The desire to "do itright" suggests a longer process. So the newproduct manager is caught between conflictingdemands of time efficiency and project effective-ness. Parallel processing compresses the develop-ment cycle without sacrificing quality.

The sequential analogy is that of a relay race:One runner, perhaps the product manager, runswith the baton for a while, passing it to the nextrunner, likely R&D. He takes over the project andruns with the baton, passing it on to production,who throws it over the wall to marketing, who, ifnot busy on more pressing matters, carries thebaton across the finish line and into the market-place. Phrases such as "handoff," "passing theproject on," "dropping the ball," and "throwing itover the wall" are common in the too many firmsthat have adopted this approach. Besides all themiscues, the process takes far too long.

In parallel processing, many activities areundertaken concurrently rather than in series,The situation is more like a rugby football gamethan a relay race, A team (not a single runner)appears on the field, A scrum or huddle ensues,after which the ball emerges. Players run downthe field in parallel, passing the ball laterally.After 25 yards or .so, the players converge foranother scrum or gate, and another stage of ac-tivities takes place. The play is far more intensethan a relay, more work gets done in a givenelapsed time period, and many players are in-volved at any one point in time.

Parallel processing, an integral facet of stage-gate systems, meansthat more activitiesoccur in an elapsedperiod of time, whichresults in time com-pression. The processis obviously morecomplex than theseries approach, re-quires more carefulmanagement, andhence points to theneed for a thoughtfulgame plan, A secondbenefit of parallelproce,ssing is multidis-ciplinary inputs (theteam passing theball). Parallel process-

ing means multifunctional, multidisciplinary in-putsdifferent activities in different parts of the

^'Estoblished criterioensure that oii projectsare evaluated consis-tently and fairly, andthat gut decisions andhidden criteria take aback seat to specifiedcriteria and thoughtfuldecisions. "

firm being undertaken concurrentiy, but all con-verging at the next scrum or gate.

Better Project Evaluations

New product resources are too valuable andscarce to misallocate. Sound project evaluations,where Go/Kill and prioritization decisions aremade, are critical to the proper allocation of de-velopment resources. But in too many firms,project evaluations are weak, deficient, or evennonexistent. In one study (Cooper and Klein-schmidt 1986), initial screening was one of themost poorly handled activities of the entire newproduct process. Further, 37 percent of projectsdid not receive a business/financial analysis priorto the R&D phase, and 65 percent of projects didnot include a pre-commercialization businessanalysis.

Typically, at the idea stage, projects wereinitiated on the basis of relatively little informa-tion and no formal criteria. When underway,projects became express trains; once the projectgot a head of steam, very little could stop it. Sub-sequent evaluations tended to be viewed as tib-stacles to be overcome, and rarely was a projectkilled once it made it to the product developmentphase. The train slowed down at the stations, butlittle stood in the way of its ultimate destination(market launch).

Most projects are unfit for commercialization,Booz, Allen & Hamilton (1982) reported that forevery seven projects at the idea stage, only onebecomes a commercial success. Effective projectevaluation is critical to successful new productprograms. Good evaluations prevent "losers"from proceeding too far, with the resulting misal-location of scarce resources. And good evalu-ations focus the resources on potential winners.

Stage-gate systems are designed to overcomethese deficiencies in project evaluation. First,project evaluations and bailout points are builtinto the process via preset gates, A project cannotpass into the next phase or .stage until the evalu-ation is done and the gate is opened. For ex-ample, in Figure 2, the decision to "move into afull-scale development project" cannot be takenuntil the Gate 3 criteria have been met. Somefirms impose a maximum dollar or time limit perstage, which if exceeded signals another gatereview.

Second, gates are characterized by a list ofpreestablished criteria. In the early gates, thesecriteria tend to be largely qualitative and dealwith "mast meet" and "sliould meet" issues. Atlater gates, a stronger financial orientation is in-troduced, where expected financial return be-comes the focus, E,stab]ished criteria ensure thatall projects are evaluated consistently and fairly,and that gut decisions and hidden criteria take a

Business Horizons / May-June 1990

back seat to specified criteria and thoughtfuldecisions. Established criteria also force a discus-sion of important issues, ensuring that none areoverlooked in the evaluation procedure.

Finally, gates provide for top managementinvolvement. Typically the gatekeepers includethe senior managers in the business unit. Sincethey have the authority t( make spending deci-sions, the project leader can seek an immediateGo/Kill/Hold decision. More importantly, he orshe can obtain immediate approval of the neededresources. Gatekeepers also have more andbroader experience than most project leaders andbring useful insights to the project. One key roleof gatekeepers, for example, is to help the proj-ect leader chait the project s path. Finally, bymanning the gates and taking a direct hand inproject evaluation from Gate 1 onward, seniormanagers "buy into" projects,

A Visible Road Map

Stage-gate systems provide a road map for theproject leader and team. The project team mem-bers, often from different functions and kjcationswithin the firm, now have a clearer idea of wherethe project stands, where it is going, and whatneeds to be done next, "At lea.st they're readingfrom the same page of the same book," com-mented a senior Exxon Chemical executive abouthis marketing and technical people as they im-plemented a stage-gate system,

Stage-gate approaches lay out the suggestedactivities for each stage of the process. None ofthese activities is mandatoryeach project isuniquebut now the project leader has a goodsense of what activities seem reasonable to con-sider f(ir the next stage of the prciject. For ex-ample, some of the actions that are sugge.sted inthe gateways model for Stage 2 in Figure 2, thedetailed investigation, include:

Undertaking a market study to identify userneeds and wants, their "must have" and "wouldlike to have" features in the product;

Undertaking a detailed technical appraisalto determine the probable technical solution, thelikelihood of completion, the co,sts and times todevelop, and the potential killer variables;

Initiating an inve.stigation through the legaldepartment to determine the copyright/patent/legal status of the project and how any road-block.-i/problems will be solved.

These suggested activities are laid out inconsiderably more detail in companies' individualgame plans.

Gates aLso provide a set of objectives for theproject leader. The prespecified deliverables orinputs lo each gate become the objectives for thenext time period. For example, the required in-puts to Gate 2 in Figure 2 might include:

"Stage-gate systemsprovide an overview ofthe entire new productprocess for senior man-agers giving structureand a vocabuiary forbetter managementand controi. "

Market analysis: size; growth; segmenta-tion; trends;

Competitive analysis: players; marketshares; strategies;

Customer reaction: reaction to concept;price sensitivities;

Development appraisal: feasibility; route;times and costs;

Production appraisal: feasibility; route;times and costs;

Legal: initial assessment from legal; Financial: payback period.With a clearly de-

fined set of objec-tives, the leader andteam are more likelyto reach the desireddestination: a well-thought-out, carefullyexecuted new prod-uct.

Finally, the out-puts of each gatehelp to define theproject and guide theproject leader. Onerole of the gatekeep-ers is to review andapprove the actionplan at each gate.Here, senior managers provide suggestions andgi.iidance to the project leader and help move theproject forward.

Senior managers also benefit from tlie visibleroad map that gateways systems provide. Execu-tives are often concerned about questions suchas, "What s the status of Project X?" or "How-many projects are coming into the launch phasein the next quarter?" Stage-gate .systems providean overview of the entire new product processfor senior managers giving stmcture and a vo-cabulary for better management and control.

One Du Pont division that recently imple-mented a stage-gate model developed a com-puter tracking system for new procluct projects,based on the gates and stages of the model. Nowan executive can review on a terminal whereProject X stands, or which projects are approach-ing Gate 2 this month. Eventually, the system willinclude a complete history on each project, soinstant information is available on any develop-ment projectincluding, for example, how theproject was rated on various criteria at each ofthe previous gate reviews.

A TYPICAL STAGE-GATE SYSTEM

T he system shown in Figure 2 is fairly ge-neric, based on one used in a typicalmanufacturing firm. But it serves as aStage-Gate Systems: A New TKI for Managing New Products 51

sample or skeleton from which to develop a cus-tom-tailored model. (Very similar stage-gate ap-proaches have been implemented in a wide vari-ety of industries, including chemicals, financialservices, and consumer nondurables). The vari-ous stages and gates are described below.

Idea

The new product process is initiated by a newproduct idea, which is submitted to Gate 1, InitialScreen.

Gate 1: Initial Screen

Initial screening is the first decision to commitresources to the project: the project is bom at thispoint. If the decision is Go, the project movesinto the Preliminary Assessment stage. Thus Gate1 signals a preliminary but tentative commitmentto the project: a flickering green light.

Gate 1 is a "gentle" screen, and amounts tosubjecting the project to a handful of key "mustmeet" and \should meet" criteria. These criteriadeal with strategic alignment, project feasibility,magnitude of the opportunity, differential advan-tage, synergy with the firm's core business andresources, and market attractiveness. Financialcriteria are not part of this first screen. A checklistfor the "must meet" criteria and a scoring model(weighted rating scales) for the "should meet"criteria are used to help focus the discussion andrank projects in this early screen.

Stage 1: Preliminary Assessment

This first and inexpensive stage has the objectiveof determining the project's technical and market-place merits. A preliminary market assessment isone facet of Stage 1 and involves a variety ofrelatively inexpensive activities: a library search,contacts with key users, focus groups, and even aquick concept test with a handful of potentialusers. The purpose is to determine market size,market potential, and likely market acceptance.

Concurrently, a preliminary technical assess-ment is carried out, involving a quick and pre-liminary in-house appraisal of the propose!product, The purpose is to assess developmentand manufacturing feasibility, and possible costsand times to execute.

Stage 1 thus provides for the gathering ofboth market and technical informationat lowcost and in a short time, so the project can bereevaluated more thoroughly at Gate 2.

Gate 2: Second Screen

This gate is essentially a repeat of Gate 1 : Theproject is reevaiuated, but in the light of the new

information obtained in Stage 1. If the decision isGo at this point, the project moves into a heavierspending stage.

At Gate 2, the project is again subjected tothe original set of "must meet" and "should meet"criteria used at Gate 1. Here, additional "shouldmeet" criteria are considered, dealing with salesforce and customer reaction to the proposedproduct, the result of new data from Stage 1.Again, a checklist and a scoring model facilitatethis gate decision. The financial return is assessedat Gate 2, but only by a quick and simple finan-cial calculation (for example, the payback pe-riod).

2: I>efnition

This is the final stage prior to product develop-ment. It is the stage that must verify the attrac-tiveness of the project prior to heavy spending.And it is the stage where the project must beclearly defined.

Here, market research studies are undertakento determine the customer's needs, wants andpreferencesthat is, to help define tlie "winning"new product. Competitive analysis is also a partof this stage. Another market activity is concepttesting, where the likely customer acceptance ofthe new product is determined.

At Stage 2, a detailed technical appraisal mustfocus on the 'do-ability" of the project. That is,customer needs and "wish lists" must be trans-lated into technically and economically feasiblesolutions. This might even involve some prelimi-nary design or laboratory work, but it should notbe construed as a full-fledged development proj-ect. An operations appraisal can I>e a part ofStage 2, where issues of manufacturability, coststo manufacture, and investment required areinvestigated. If appropriate, detailed legal/patent/copyright work is undertaken.

Finally, a detailed financial analysis is con-ducted as an input to Gate 3. This financial analy-sis typically involves a discounted cash flow ap-proach, complete with sensitivity analysis.

Gate 3: Decision on Business Case

This is the final gate prior to the DevelopmentStage, the last point at which the project can bekilled before entering heavy spending. Once pastGate 3, financial commitments are substantial. Ineffect. Gate 3 means "go to a heavy spend."

The project is once again subjected to the setof "must meet" and "should meet" criteria used atGate 2. Next, the qualitative side of this evalu-ation involves a review of each of the activities inStage 2, checking that the activities were under-taken, the quality of execution was sound, andthe results were positive. Finally, because a

52 Business Horizons / May-June 199

heavy spending commitment is the result of a Godecision at Gate 3, the results of the financialanalysis are an important part of this screen.

A .second part of Gate 3 concerns definitionof the project. At Gate 3, agreement must bereached on a number of key items before theproject proceeds into the Development Stage.These items include target market definition;definition of the product concept, specification ofa product pi)sitioning strategy, and delineation ofthe protluct benefits to be delivered; and agree-ment on essential and desired product features,attributes, and specifications.

Plans that chart the path forwardthe devel-opment plan and the preliminary operations andmarketing plansare reviewed and approved atthis gate.

Stage 3: Development

Stage 3 involves the development of the productand (concurrently) of detailed test, marketing,and operations plans. An updated financial analy-sis is prepared, and legal/patent/copyright issuesare resolved.

Gate 4: Post-Development Review

The Post-Development Review is a check on theprogress and the continued attractiveness of theproduct and pniject. Development work is re-viewed and checked, ensuring that the work hasbeen completed in a quality fashion. This gaterevisits the economic que.stion via a revised fi-nancial analysis based on new and more accuratedata. The te.st or validation plans for the nextstage are approved for immediate implementa-tion, and the detailed marketing and operations[lans are reviewed for probable future execution.

Stage 4: Validation

This stage tests the entire viability of the project:the product itself; the production process; cus-ttimer acceptance; and the economics of the proj-ect. A number of activities are undertaken atStage 4:

In-house product tests; to check on prod-uct quality and product performance;

U.ser or field trials of the product: to verifythat the product functions under actual use con-ditions, and also to gauge potential customers'reaction to the product;

Trial or pilot production: to test and debugilie production proce.ss, and to determine moreprecise production co.sts and rates;

Pretest market, test market, or trial sell: to.L auge customer reaction, measure the effective-ness of the launch plan, and determine expectedmarket share and revenues;

Revised financial analysis; to check on thecontinued economic viability of the project,based on new and more accurate revenue andcost data.

Gate 5: Pre-Commercialization Decision

This final gate opens the door to full commer-cialization. It is the final point at which the proj-ect can still be killed. This gate focuses on thequality of the activities at the Validation Stage andtheir results. Financial projections play a key rolein the decision to move ahead. Finally, the opera-tions and marketing plans are reviewed and ap-proved for implementation in Stage 5.

Stage 5: Commercialization

This final stage involves implementation of boththe marketing launch plan and the operationsplan.

Post-Implementation Review

At some point following commercialization, thenew product project must be terminated. Theteam is disbanded, and the product becomes a"regular product" in the firm's line. This is alsothe point where the project and product's per-formance is reviewed. The latest data on reve-nues, costs, expenditures, profits, and timing arecompared to projections to gauge performance.Finally a post-audita critical assessment of theproject's strengths and weaknesses, what we canlearn from this project, and how we can do thenext one betteris carried out. This reviewmarks the end of the project.

Not all projects pass through every stage ofthe model. Usually there are at least two or threecategories of projectsranging from sales devel-opments (small projects in response to a singlecustomer request) to major projects involvingheavy expenditures. Standard definitions of proj-ect types are developed and based on projectscope and investment required. Appropriateroutes are determined for each type of project.The routing for any project at the idea stage isdecided in the first gate. Initial Screen. Projectsthat are "closer to home," hence less risky, arenecessarily treated in a less rigorous fashion than"step out" projects, major projects with high risk.

T he basic benefits of the stage-gate processare evident. The model puts disciplineinto a process that, in too many firms, isad hoc and seriously deficient. The process isvisible and relatively simple: what is required ateach stage and gate is understood by all. Theprocess provides a road map to facilitate theproject, and it better defines the project leader's

Syslemsi A New Tool for Managing New Products 53

objectives and tasks, Also, the process builds inevaluation stages to better rank projects and fo-cus resources.

Parallel processing is a facet of stage-gatemodels, ensuring timely completion of projects.Moreover, stage gate systems build in key activi-ties and focus attention on often-deficient areassuch as the predevelopment and market-orientedtasks. And finally, there is a quality focus to theprocess: a complete process with quality of exe-cution.

Product innovation will always be a high-riskendeavor. The stage-gate system is merely a disci-pline that builds the success ingredients into theinnovation process by design rather than bychance. The results are better decisions, morefocus, fewer failures, and faster developments. O

References

Booz, Allen & Hamilton, New Product Management forthe 1980s (New York: Booz, Allen & Hiunilton, 1982).

Coopers & Lybrand Consulting Group, Business Plan-ning in the FAghties: Tl.ie New Marketing Shape of NorthAmerican Corporations (New York: Coopers &LybrandA'ankelovich, Skelly & White, 1985).

R.G. Cooper, "Why New Industrial Products Fail,"Industrial Marketing Management. 4, 1975, pp. 315-326.

R.G. Cooper. Winning at New Products iReai^g,Mass,; Addison-Wesley, 1986).

R.G. Cooper and E.J. Kieinschmidt, "An Investigationinto the New Product Proce.s.s: Steps. Dtrficiendes andImpact," Journal of Product Innovation Management,3(2), 1986, pp. 71-85.

R.G. Cooper and H.J. Kleiaschmidt, "Resource Alloca-tion in the New Product Proces.s," Industrial MarketingManagement, 17(3), 1988, pp. 249-262,

CM. Crawford, "New Product Failure RatesFacts andFallacies," Research Management, September 1979, pp.9-13.

D.S. Hopkins, "New Products Winners and Losers,"ITie Conference Board, Report no, 773, New York,1980.

D.S. Hopkins and E.L. Bailey, "New Procluct Pressures,"Conference Board Record S(6), 1971, pp. 16-24.

Bro Uttal, "Speeding New Ideas to Market," Fortune,March 1987, pp, 62-66.

Robert G. Cooper is a professor of mar-keting ot McMaster University, Hamilton,Ontario.

54 Business Horizons / May-June 1990