Lean: The mother of invention?

We all know the story of the spread of lean principles. Starting in Japan and spreading around the world, lean thinking and the values of the Toyota Production System have revolutionised manufacturing.

With the appearance of, Lean Thinking, this revolution expanded beyond manufacturing. But alas, the world isn’t always friendly towards new ideas and when lean methods first arrived in product development the reception was not warm.

Unenlightened engineers grimaced and complained, “You don’t understand. That won’t work here; we are not manufacturing.” And in fact, although some of the grimacing was due to the ineptness of the initial efforts, the engineers were right, inept? Isn’t that a bit harsh? I’ll let you be the judge, but I saw lean consultants visit engineering departments and draw outlines on desks to mark the location of staplers. This 5S approach brought great value to manufacturing, but being able to find staplers quickly had no meaningful impact on product development performance and so the use of this tool was flawed.

Context is everything
Such attempts to impose lean practices in product development environments did not work because they assumed leverage for value creation would be found in the same places that it was found in manufacturing.

It is not. The way product development makes money is fundamentally different from the way manufacturing makes money. Understanding what drives economic results in any environment is the first step to improving economic performance.

One key difference between manufacturing and product development has huge implications: manufacturing is repetitive; product development never covers the same ground twice. To consistently produce an identical product a thousand times in manufacturing will make you money. In design replication it is 100% waste.

This introduces a need to constantly change the recipe of day-to-day work and change introduces uncertainty. It is self-evident that variability is a negative for manufacturing, but it is a bad assumption that what is negative in one environment will be negative in another. In product development variability is the companion of innovation, and innovation is the engine of profit. Taking this into consideration it shouldn’t surprise you that the early efforts to eliminate product development variability had some rather embarrassing side effects.

Learning from mistakes
Today we have progressed to what we call, Second Generation Lean Product Development. While many lean manufacturing ideas require absolutely no modification for use in product development, it is my belief that mixed with these good ideas are others that misbehave when they are taken away from a manufacturing setting. These non-transportable ideas are a great danger and must be left behind.

Starting with the positives, it is clear that reducing in-process inventory shortens cycle time. Reducing batch size is the key to reducing in-process inventory. Reducing transaction cost enables batch size reduction. These are solid ideas which can be applied to the product development process for great advantage.

For the consideration of our non-transportable ideas however, let’s return to the single most toxic idea for product development, the idea that variability is bad. No lean manufacturing expert would question this belief; their experience repeatedly proves to them that it is true and this raises a big challenge for companies with a deep understanding of lean manufacturing to unlearn some of their core beliefs.

The intrinsically high variability of lean product development, and its beneficial influence on innovation, motivates product developers to be clever in handling variability. This can be achieved by using queuing theory, a branch of applied statistics dating back to 1909. Queuing theory lies at the core of the design of modern telecommunications systems, and it gives us deep insights into how to achieve flow in the presence of variability. In fact, the more we understand queueing theory the more we realise that telecommunications systems are probably more useful role models for product developers than manufacturing systems.

Another idea from lean manufacturing requiring major overhaul for lean product development is the concept of first-in, first-out or FIFO. This is optimal whenever all work has the same delay cost and the same task duration. But, what happens when we deal with non-homogeneous work – as in product development? Just imagine a hospital emergency room that decided to use a FIFO processing. A patient with a heart attack would be told to wait while a patient with an earache was treated. FIFO is fundamentally wrong for non-homogeneous flows.

Other prioritisation approaches make much more sense. For example, in product development it is often optimum to prioritise work on the basis of cost of delay.

This issue of costs brings us on to the lean manufacturing belief that it is better to prevent a problem than it is to correct it. This is the foundation for our drive to minimise defects and to achieve 6 sigma quality levels. Is this appropriate in product development?

Manufacturing environments favour a focus on prevention because they have long run lengths.

This means the cost of avoiding a problem is paid back hundreds or thousands of times. Product development is often a one-time activity – this makes it much harder to recover the effort invested in prevention. To use a simple analogy, most of us use the spellcheckers in our word processors. If prevention was really a universally superior strategy we should teach ourselves to become perfect typists and perfect spellers. This makes no sense because the cost of prevention far exceeds the cost of correction. In fact, whenever the cost of prevention is higher than the cost of correction it makes no economic sense to prevent problems. Such ideas are sacrilege to lean manufacturing experts.

Sacred lean manufacturing concepts like “pull” also break down in product development, where we must anticipate demand. Why? Product development lead times are much longer than acceptable customer response times. To prevent customers from waiting we must start working on a product long before customers ask for it, or risk being late to market. Imagine a baker using “pull” and waiting for a customer to order a baguette before they started baking it.

And this brings us to the most fundamental difference between the approach of lean product development and that of lean manufacturing.

Lean manufacturing is, at its heart, a qualitative system. Its rules are treated as tenants of faith. It is a world of ‘always’ and ‘never’. Always eliminate variability, always let the customer pull, always front-load decisions.

In lean product development there are no absolutes and this requires quantitative rather than qualitative approaches. Should we operate the CAD areas at 80% utilisation with a two week queue, or 90% utilisation with a four week queue? We can’t compare a two week difference in queue size with a 10% difference in utilisation unless we can express them both in the same unit of measurement. In practice, we evaluate such decisions by translating all performance changes into their impact on lifecycle profits. Rigorous economic decision making is one of the key hallmarks of second generation lean product development.

Be bold
Lean product development is an exciting body of knowledge and probably the only management approach that can simultaneously improve cycle time, quality, and efficiency. Furthermore lean product development is rapidly evolving and therefore a challenge for companies that don’t like using ideas until they have become stable. Yet, we should recognise that there is a difference between stability and usefulness. The original IBM personal computer was introduced in 1983. If you decided to wait for it to stop improving before you bought one, then would you still be without a computer today. Consider the implications for competitive wellbeing.