Upgrade Your New-Product Machine: This Process Allows the Product Team to Prepare a Solid Business Case before Beginning the Costly Development Stage

Most manufacturing companies have two product machines: the "old-product machine" makes products we know customers want, while the "new-product machine" makes product designs we think customers want. The old-product machine is the manufacturing process. We call its output "old" simply because we've made these products for customers before and they can no longer be considered new. The new-product machine is the new-product development process. Its output is less tangible but at least as critical: a product design that is attractive to customers and profitable to the supplier. The two machines are linked in that the output of the new-product machine is used to establish the raw materials, settings and output specifications for the old-product machine. The old-product machine is a decent machine. After a great deal of attention by management and employees with belts of various colors, some companies have had impressive Six Sigma success, producing three or fewer defects per million attempts. Then there is the new-product machine. We know that only one in four projects becomes a commercial success after the decision has been made to enter the costly development stage (1,2). So, as shown in Figure 1, this machine is producing three defects per four attempts. And the attempts are serious, since most companies now conduct gate reviews prior to the development stage (3). These defects squander immense R&D and commercial resources, as well as fill up the old-product machine's capacity in making low-margin output. In other words, nearly every facet of the company--R&D, marketing, sales and manufacturing--is affected. To be fair, the new-product machine has a more difficult assignment. Just as we prize an original painting over a reprint--because we recognize the creativity and effort required--so we appreciate the challenge of developing a winning new-product design. We measure the success of the old-product machine with the question, "Is the out-put the same?" while we ask of the new-product machine, "Is the output different--and better?" Also, because the new-product machine deals in intangibles, we find its variables much more difficult to measure and control. [FIGURE 1 OMITTED] But can we do no better? If we could reach just "Two Sigma," the average company's new-product success rate (from the development stage) would skyrocket to about two in three attempts. Your chief technology officer might not choose to have Two Sigma celebrations, but there is more latent corporate value that can be unlocked here than anywhere else. As you consider the millions your company invests in R&D alone, can you think of any other area where such waste is tolerated? So, how many of these defects are part and parcel of new product development, and how many can be driven out--in much the way variability is relentlessly pursued in the old-product machine? Fortunately, we have known for some time which parts of the new-product machine are malfunctioning, and we even know some of the upgrades that will help. Let's look at the malfunctions first. Five Valuation Errors Imagine you are at a gate review trying to decide if a project should enter the development stage. You are about to make a valuation--an estimation of something's worth (4)--weighing the likely benefits of this project against its costs. The golden rule of valuation is: Make your decision when you have gathered the most facts and spent the least money. Since pre-development activities are focused on gathering facts and most companies make at least 90 percent of their project investment after this point (5), you are at the right meeting. So far, so good. Now, you and your colleagues must avoid five types of valuation errors (see Figure 2). We have discovered most of these errors the hard way: by making them personally as we led product development teams and then observing these same errors in dozens of businesses at our employers and consulting clients. …