Tuesday, 18 December 2001

Reflections on Development

Software Development is just that ‘Development’. Writing software involves creating something new. (After all, if developing software were repeatable, a computer could be programmed to do it.)

If we cast around for other uses of the word ‘Development’, we find that it means:
  1. Fund raising for a non-profit corporation

  2. Turning vacant land into a subdivision or shopping center

  3. Commercializing new products

Bypassing definitions (1) and (2) for the moment, let’s consider the third. If we determine good New Product Development practices, might we come up with some good ideas for Software Development? It seems that Software Development might learn a lesson or two from New Product Development.

New Product Development has been around for a long time. Some companies do a great job of it, and we can look at corporate track records and find out which ones do it well. In fact, some companies do New Product Development so well that they are considered benchmarks of excellent Product Development practices. Two companies which clearly qualify for this honor are Toyota and 3M.

Toyota has the fastest product development times in the automotive industry, is a consistent leader in quality, has a large variety or products designed by a lean engineering staff, and has maintained US market share despite a strong yen.

Almost half of all 3M products are only 5 years old, and its innovation machine continues to roll out an astounding number high quality of new products every year. New product development practices at 3M have been studied and imitated for over a decade.

Development methods used by 3M and Toyota may be distilled into to four key practices:
  1. Use Entrepreneurial Teams

  2. Focus on Business Value

  3. Overlap Development Phases

  4. Converge on Solutions

Use Entrepreneurial Teams
3M has a 15% rule: All scientists may use 15% of their time working on any project they wish. In addition, the company expects all new product teams to be led by a ‘Product Champion’. The champion typically has an entrepreneurial vision of a new product, and recruits scientists to use their 15% time to help develop the vision. New product teams can make great progress on company time with no official approval, and once they have developed their idea far enough, they appeal to the company for funding to bring their venture to market.

A new product development team at 3M is cross-functional, collaborative, autonomous and self-organizing. It deals well with ambiguity, accepts change, takes initiative and assumes risks. If the team is making progress toward a new product, it will be left alone.

Teams at Toyota are led by a chief engineer who is expected to understand the market and whose primary job is vehicle system design. The Chief Engineer is totally responsible for vehicle development, similar to the product champion at 3M. This creates an entrepreneurial environment in which the second practice can be accomplished:

Focus on Business Value
3M new product development teams have a wide measure of freedom and extremely challenging goals. But these goals are not cost, schedule and scope. At each review, the key hurdle the team must pass is proving that the proposed product has a good margin potential. One reason why 3M gets hundreds of high margin new products each year is because senior management decision-making focuses teams on this broad strategic goal.

Toyota was the birthplace of Lean Production, originally known as the Toyota Production System (TPS). The fundamental concept of the is to eliminate waste, which means eliminating anything which does not add business value. Toyota was the first company to eliminate sub-optimized measurements and allow overall business value to drive decisions at every level. This tends to knock down barriers between departments and facilitates the third practice:

Overlap Development Phases
At 3M, product development appears to move through a set of phases: concept, feasibility, product and process development, pilot production, scale-up. In fact, with a single cross-functional team handling the product from start to finish, a new product will be manufactured on a pilot line and test marketed as early as possible. Changes will be made to the product spec, to the manufacturing process and to the product design, based on feedback from the manufacturing line and from customers. The major problem with this approach is that when a pilot product is successful, the team may not be able to scale up fast enough to meet demand. It’s a nice problem to have.

Toyota uses compressed overlapping phases to rapidly bring new products to market. Traditional serial engineering in the automotive industry involves moving a design from styling to marketing to body to chassis to manufacturing. At Toyota, concurrent engineering takes place: styling, body, chassis and manufacturing designs are done simultaneously, even as market investigation continues. This works because each function develops multiple options, and the Chief Engineer integrates the product by looking for intersections between feasible regions of functional designs. Which leads to the fourth practice:

Converge on Solutions
Toyota makes more clay models than other companies, freezes the nominal dimensions far later than US manufacturers, and obtains detailed input on process capability at the earliest stages of design. Instead of optimizing from a single starting point, tradeoffs are made by rigorously evaluating sets of alternatives, bringing forward multiple options, making tradeoffs based on data, and deciding as late as possible in the design process.

Similarly, 3M teams would not be expected to know at the beginning of a new product development project exactly what they will commercialize. After all, Art Fry was trying to make a sticky bookmark when he invented Post-it® Notes. When there are inventions to be made and markets to be tested, the product development team tries out many ideas, abandoning the ones that fail and pursuing those that work. Failure is not a problem at 3M – less than half of new product development projects are expected to succeed. But from an abundance of options comes an large number of innovative products, including a regular stream of blockbuster successes.

Conclusion
Software Development, in as much as it is really ‘Development’, might well borrow a few ideas from companies that have mature, high performance development practices. Truly innovative development comes from entrepreneurial teams expected to deliver business value, where all functions work in parallel to converge on an optimal solution.

Screen Beans Art, © A Bit Better Corporation