Based on the paper by Wheelwright and Clark (1992) and the case information, prepare illustration showing where each project fits in a Derivative,

1. Based on the paper by Wheelwright and Clark (1992) and the case information, prepare illustration showing where each project fits in a Derivative, Platform, Breakthrough, and R&D classification system.

Creating Project Plans to Focus Product Development

by Steven C. Wheelwright and Kim B. Clark

HarvardBusinessReview

Reprint 92210

With an "aggregate project plan," companies map out and manage a set

Creating Project Plans to Focus

by Steven C. Wheelwright and Kim B. Clark

The long-term competitiveness of any manufac- turing company depends ultimately on the success of its product development capabilities. New prod- uct development holds hope for improving market position and financial performance, creating new industry standards and new niche markets, and even renewing the organization. Yet few develop- ment projects fully deliver on their early promises. The fact is, much can and does go wrong during de- velopment. In some instances, poor leadership or the absence of essential skills is to blame. But often problems arise from the way companies approach the development process. They lack what we call an "aggregate project plan."

Consider the case of a large scientific instru- ments company we will call PreQuip. In mid-1989, senior management became alarmed about a rash of late product development projects. For some months, the development budget had been rising even as the number of completed projects declined. And many of the projects in the development pipeline no longer seemed to reflect the needs of the market. Management was especially troubled be-

cause it had believed its annual business plan pro- vided the guidance that the marketing and engi- neering departments needed to generate and sched- ule projects.

To get to the root of the problem, the chief execu- tive first asked senior managers to compile a list of all the current development projects. They discov- ered that 30 projects were under way-far more than anticipated, and, they suspected, far more than the organization could support. Further analysis re- vealed that the company had two to three times more development work than it was capable of completing over its three-year development plan- ning horizon. (See the chart "PreQuip's Develop- ment Predicament: Overcommitted Resources.")

With such a strain on resources, delays were in- evitable. When a project ran into trouble, engineers

Steven C. Wheelwright and Kim B. Clark are both profes- sors of business administration at the Harvard Business School, where they teach technology and operations management. This article is taken from their book, Rev- olutionizing Product Development: Quantum Leaps in Speed, Efficiency, and Quality, published by The Free Press in April 1992.

Copyright © 1992 by the Presidents and Fellows of Harvard College. All rights reserved. DRAWING BY MARK STEELE

of strategic development projects.

Product Development

from other projects were reassigned or, more com- monly, asked to add the crisis project to their al- ready long list of active projects. The more projects they added, the more their productivity dropped. The reshuffling caused delays in other projects, and the effects cascaded. Furthermore, as deadlines slipped and development costs rose, project man- agers faced pressure to cut corners and compromise quality just to keep their projects moving forward.

The senior management team also discovered that the majority of PreQuip's development re- sources-primarily engineers and support staff-was not focused on the projects most critical to the business. When questioned, project leaders admit- ted that the strategic objectives outlined in the an- nual business plan had little bearing on project se- lection. Instead, they chose projects because engineers found the technical problems challeng- ing or because customers or the marketing depart- ment requested them. PreQuip had no formal pro- cess for choosing among development projects. As long as there was money in the budget or the person making the request had sufficient clout, the head of

the development department had no option but to accept additional project requests.

Many engineers were not only working on non- critical projects but also spending as much as 50% of their time on nonproject-related work. They re- sponded to requests from manufacturing for help with problems on previous products, from field sales for help with customer problems, from quali- ty assurance for help with reliability problems, and from purchasing for help with qualifying vendors. In addition to spending considerable time fixing problems on previously introduced products, engi- neers spent many hours in "information" and "up- date" meetings. In short, they spent too little time developing the right new products, experimenting with new technologies, or addressing new markets.

PreQuip's story is hardly unique. Most organiza- tions we are familiar with spend their time putting out fires and pursuing projects aimed at catching up to their competitors. They have far too many pro- jects going at once and all too often seriously over- commit their development resources. They spend too much time dealing with short-term pressures

HARVARD BUSINESS REVIEW March-April 1992

3

PRODUCT DEVELOPMENT

Total Available Capacity

960

460

2783 months

2956 months

2178 months

1989 1990 1991

Nonproject-related engineering time Project engineering time (for 30 projects)

PreQuip's Development Predicament: Overcommitted Resources

PreQuip had 960 engineering months each year to allocate to development work. But combining the time it would take to keep its current 30 projects on schedule with the time engineers spent doing nonproject development work, the company found it had overcommitted its development resources for the next three years by a factor of three.

and not enough time on the strategic mission of product development.

Indeed, in most organizations, management di- rects all its attention to individual projects-it micromanages project development. But no single project defines a company's future or its market growth over time; the "set" of projects does. Com- panies need to devote more attention to managing the set and mix of projects. In particular, they should focus on how resources are allocated be- tween projects. Management must plan how the project set evolves over time, which new projects get added when, and what role each project should play in the overall development effort.

The aggregate project plan addresses all of these issues. To create plan, management categorizes projects based on the amount of resources they con- sume and on how they will contribute to the com- pany's product line. Then, by mapping the project types, management can see where gaps exist in the development strategy and make more informed de- cisions about what types of projects to add and when to add them. Sequencing projects carefully, in turn, gives management greater control of resource

allocation and utilization. The project map also reveals where development capabilities need to be strong. Over time, companies can focus on adding critical re- sources and on developing the skills of individual contributors, project leaders, and teams.

Finally, an aggregate plan will enable management to improve the way it manages the develop- ment function. Simply adding projects to the active list-a com- mon practice at many compa- nies-endangers the long-term health of the development pro- cess. Management needs to cre- ate a set of projects that is con- sistent with the company's development strategies rather than selecting individual projects from a long list of ad hoc propos- als. And management must be- come involved in the develop- ment process before projects get started, even before they are fully defined. It is not appropriate to give one department-say, engi- neering or marketing-sole re- sponsibility for initiating all pro- jects because it is usually not in a

position to determine every project's strategic worth.

Indeed, most companies-including PreQuip- should start the reformation process by eliminating or postponing the lion's share of their existing pro- jects, eventually supplanting them with a new set of projects that fits the business strategy and the ca- pacity constraints. The aggregate project plan pro- vides a framework for addressing this difficult task.

How to Map Projects

The first step in creating an aggregate project plan is to define and map the different types of develop- ment projects; defining projects by type provides useful information about how resources should be allocated. The two dimensions we have found most useful for classifying are the degree of change in the product and the degree of change in the manufac- turing process. The greater the change along either dimension, the more resources are needed.

Using this construct, we have divided projects into five types. The first three-derivative, break- through, and platform-are commercial develop-

4

HARVARD BUSINESS REVIEW March-April 1992

Engineering Months

ment projects. The remaining two categories are re- search and development, which is the precursor to commercial development, and alliances and part- nerships, which can be either commercial or basic research. (See the chart "Mapping the Five Types of Development Projects.")

Each of the five project types requires a unique combination of development resources and manage- ment styles. Understanding how the categories dif- fer helps managers predict the distribution of re- sources accurately and allows for better planning and sequencing of projects over time. Here is a brief description of each category.

Derivative projects range from cost-reduced ver- sions of existing products to add-ons or enhance- ments for an existing production process. For exam- ple, Kodak's wide-angle, single-use 35mm camera, the Stretch, was derived from the no-frills Fun Saver introduced in 1990. Designing the Stretch was primarily a matter of changing the lens.

Development work on derivative projects typi- cally falls into three categories: incremental prod- uct changes, say, new packaging or a new feature, with little or no manufacturing process change; in- cremental process changes, like a lower cost manu- facturing process, improved reliability, or a minor change in materials used, with little or no product change; and incremental changes on both dimen- sions. Because design changes are usually minor, incremental projects typically are more clearly bounded and require substantially

fewer development resources than the other categories. And because derivative projects are completed in a few months, ongoing management involvement is minimal.

Breakthrough projects are at the

other end of the development spec-

trum because they involve signifi-

cant changes to existing products

and processes. Successful break-

through projects establish core products and pro- cesses that differ fundamentally from previous gen- erations. Like compact disks and fiber-optics cable, they create whole new product category that can define a new market.

Because breakthrough products often incorporate revolutionary new technologies or materials, they usually require revolutionary manufacturing pro- cesses. Management should give development teams considerable latitude in designing new pro- cesses, rather than force them to work with existing plant and equipment, operating techniques, or sup- plier networks.

Platform projects are in the middle of the develop-HARVARD BUSINESS REVIEW March-April 1992

ment spectrum and are thus harder to define. They entail more product and/or process changes than derivatives do, but they don't introduce the untried new technologies or materials that breakthrough products do. Honda's 1990 Accord line is an exam- ple of a new platform in the auto industry: Honda introduced a number of manufacturing process and product changes but no fundamentally new tech- nologies. In the computer market, IBM's PS/2 is a personal computer platform; in consumer products, Procter & Gamble's Liquid Tide is the platform for a whole line of Tide brand products.

Well-planned and well-executed platform prod- ucts typically offer fundamental improvements in cost, quality, and performance over preceding gener- ations. They introduce improvements across a range of performance dimensions-speed, function- ality, size, weight. (Derivatives, on the other hand, usually introduce changes along only one or two di- mensions.) Platforms also represent a significantly better system solution for the customer. Because of the extent of changes involved, successful plat- forms require considerable up-front planning and the involvement of not only engineering but also marketing, manufacturing, and senior manage- ment.

Companies target new platforms to meet the needs of a core group of customers but design them for easy modification into derivatives through the addition, substitution, or removal of features. Well-

Most companies should start reforming their development process by eliminating the lion's share of existing projects.

designed platforms also provide a smooth migra- tion path between generations so neither the cus- tomer nor the distribution channel is disrupted.

Consider Intel's 80486 microprocessor, the fourth in a series. The 486 introduced a number of perfor- mance improvements; it targeted a core customer group-the high-end PC/workstation user-but vari- ations addressed the needs of other users; and with software compatibility between the 386 and the 486, the 486 provided an easy migration path for existing customers. Over the life of the 486 platform, Intel will introduce a host of derivative products, each offering some variation in speed, cost, and perfor- mance and each able to leverage the process and

5

PRODUCT DEVELOPMENT

Research and advanced development projects

More

Product Change

Next Generation Product

Less

Addition Derivatives

to Product and

Family Enhancements

New Core Product

Breakthrough projects

Platform projects

Derivative projects

New Core Process

Next Generation Process

Single Department Upgrade

Incremental Change

R&D

Alliances

and partnership projects

(can include any of the above project types)

Breakthrough

Platform

Derivative

Mapping the Five Types of Development Projects

product innovations of the original platform. Platforms offer considerable competitive lever- age and the potential to increase market penetra- tion, yet many companies systematically under- invest in them. The reasons vary, but the most common is that management lacks an awareness of the strategic value of platforms and fails to create well-thought-out platform projects. To address the problem, companies should recognize explicitly the need for platforms and develop guidelines for making them a central part of the aggregate project

plan.

Research and development is the creation of the

know-how and know-why of new materials and technologies that eventually translate into com- mercial development. Even though R&D lies out- side the boundaries of commercial development,

we include it here for two reasons: it is the precur- sor to product and process development; and, in terms of future resource allocation, employees move between basic research and commercial development. Thus R&D projects compete with commercial development projects for resources. Because R&D is a creative, high-risk process, companies have different expectations about re- sults and different strategies for funding and man- aging it than they do for commercial development. These differences can indeed be great, but a close relationship between R&D and commercial devel- opment is essential to ensure an appropriate bal- ance and a smooth conversion of ideas into prod- ucts.

Alliances and partnerships, which also lie out- side the boundaries of the development map, can be

6

HARVARD BUSINESS REVIEW March-April 1992

Less Process Change More

formed to pursue any type of project-R&D, break- through, platform, or derivative. As such, the amount and type of development resources and management attention needed for projects in this category can vary widely.

Even though partnerships are an integral part of the project development process, many companies fail to include them in their project planning. They often separate the management of

partnerships from the rest of the de-

velopment organization and fail to

provide them with enough develop-

ment resources. Even when the part-

ner company takes full responsi-

bility for a project, the acquiring

company must devote in-house

resources to monitor the project,

capture the new knowledge being

created, and prepare for the manufac-

turing and sales of the new product.

All five development categories are vital for cre- ating a development organization that is responsive to the market. Each type of project plays a different role; each requires different levels and mixes of re- sources; and each generates very different results. Relying on only one or two categories for the bulk of the development work invariably leads to subop- timal use of resources, an unbalanced product offer- ing, and eventually, a less than competitive market position.

PreQuip's Project Map

Using these five project types, PreQuip set about changing its project mix as the first step toward re- forming the product development process. It started by matching its existing project list to the five cate- gories. PreQuip's product line consisted of four kinds of analytic instruments-mass spectrometers, gas and liquid chromatographs, and data handling and processing equipment-that identified and iso- lated chemical compounds, gases, and liquids. Its customers included scientific laboratories, chemi- cal companies, and oil refineries-users that needed to measure and test accurately the purity of raw materials, intermediate by-products, and finished products.

PreQuip's management asked some very basic questions in its attempt to delineate the categories. What exactly was a breakthrough product? Would a three-dimensional graphics display constitute a breakthrough? How was a platform defined? Was a full-featured mass spectrometer considered a plat- form? How about a derivative? Was a mass spec- trometer with additional software a derivative?

None of these questions was easy to answer. But after much analysis and debate, the management team agreed on the major characteristics for each project typeand assigned most of PreQuip's 30 proj- ects to one of the five categories. The map revealed just how uneven the distribution of projects had be- come-for instance, less than 20% of the company's projects were classified as platforms. (See the chart

When the dust had settled, PreQuip had reduced the number of development projects from 30 to 11.

HARVARD BUSINESS REVIEW March-April 1992

7

"Before: PreQuip's Development Process Was Chaotic. . . . ")

Management then turned its attention to those development projects that did not fit into any cate- gory. Some projects required substantial resources but did not represent breakthroughs. Others were more complicated than derivative projects but did not fall into PreQuip's definition of platforms. While frustrating, these dilemmas opened man- agers' eyes to the fact that some projects made little strategic sense. Why spend huge amounts of money developing products that at best would produce on- ly incremental sales? The realization triggered a re- examination of PreQuip's customer needs in all product categories.

Consider mass spectrometers, instruments that identify the chemical composition of a compound. PreQuip was a top-of-the-line producer of mass spectrometers, offering a whole series of high-per- formance equipment with all the latest features but at a significant price premium. While this strategy had worked in the past, it no longer made sense in a maturing market; the evolution of mass spectrome- ter technology was predictable and well defined, and many competitors were able to offer the same capabilities, often at lower prices.

Increasingly, customers were putting greater em- phasis on price in the purchasing decision. Some customers also wanted mass spectrometers that were easier to use and modular so they could be in- tegrated into their own systems. Others demanded units with casings that could withstand harsh in- dustrial environments. Still others required faster operating speeds, additional data storage, or self- diagnostic capabilities.

Taking all these customer requirements into ac-

PRODUCT DEVELOPMENT

R&D