THE CHANGING FACE OF NEW PRODUCT DEVELOPMENT

by Jim Highsmith, Director, Agile Project Management Practice

New product development -- be it industrial products, consumer products, or internal business processes -- is being driven by two resolute forces: the continuing demand for innovation and the plunging cost of change (low-cost exploration) that is affecting a growing range of industries. As the uncertainty about the outcome of a development effort increases, and as the complexity of the interactions of design variables daunts cause-and-effect analysis, exploration through experimentation becomes the most effective and reliable mode of discovery. When we can conduct 1,000 experiments a day for $10 a piece, creating elaborate designs that take a month to complete makes no sense. However, conducting 5,000 random experiments makes no sense either. Good experiments require good experimental design.

It is hard to overstate the impact of low-cost exploration on product development. The competitive advantage that will accrue to those companies that can adjust their development and managerial processes appropriately is undeniably valuable. A significant portion of this strategy will lie in software and the manipulation of bits rather than atoms, thus driving a low-cost exploration product development strategy that strives to do the following:

• Fill products with "bits"

• Create a "bit"-oriented product development lifecycle (i.e., model and simulate products in software as far into the lifecycle as possible)

• Relentlessly drive down the cost of changing bits (low-cost iteration)

• Develop people and processes capable of the above strategies (agile people and processes)

Embedded software has rapidly become a critical piece of industrial products. Cell phones now boast a million or more lines of code. Automobiles have microprocessors for everything from fuel injection to transmission shifting. Airplanes are fly-by-wire, completely electronically controlled. The human genome sequencing is now available on large databases. An effective software development team can often create and adapt features faster than a hardware group can.

Software -- developed well, of course -- is more flexible than hardware. In fact, drastically shortening product development lifecycles forces hardware engineers to order parts earlier and earlier because of lead times, locking in the hardware designs earlier than the engineers would like. Software's flexibility is often used to correct hardware problems or add new features after hardware designs are fixed. So, in general, the more bits there are in a product, the better. As bits replace atoms, products can be developed faster, wiith more features and greater flexibility.

But you can't drive a bit, sit on a bit, or use a bit to hit a golf ball. At some point, atoms must be assembled into products that we can use. The critical question is -- when do we assemble the atoms? There is a furniture manufacturer in West Virginia that completes its entire product design process in software. At the end of the design phase, the software generates instructions to manufacturing robots, and the furniture parts are cut and assembled. The more chemical compound and biological response information available in large databases, the further into the drug development process companies can go before lengthy and expensive activities such as animal testing have to begin. Simulations, modeling, and prototyping, for a wide array of products, are all increasingly being done in software prior to any physical assembly. Once physical assembly begins, for either design testing or manufacturing, flexibility suffers.

The point is that the further into the product development process that teams can manipulate bits rather than atoms, the more effective the process will be. So dual strategies are critical: 1) increase the number of bits in products; and 2) manipulate bits rather than atoms as far into the development process as possible.

-- Jim Highsmith, Director, Agile Project Management Practice