People often search for waste in R&D using the rule of the 7 wastes. This is useful, but misses the largest opportunities to remove waste and accelerate value within your development organization. Here are 7 ideas to help you think about, find and eliminate the biggest wastes in R&D.
1. Eliminate re-learning what you already know. R&D requires learning new things, typically through experimentation. Once an experiment is complete and its answer is known, there is zero value in learning it a second time. Hence, single easiest waste to identify in R&D is re-learning something that is already known. Too many scientists and engineers spend weeks in the laboratory proving known principles and designs.
2. “Failure” is not a waste. The purpose of R&D is to develop new products and processes that create customer value. With that in mind, it is useful to recognize that nobody knows what will prove valuable to a customer until an invention is on the market. “Failed” experiments are part of learning what works and what does not work, what is valued and what is not valued. Given some caveats, failed experiments are highly value added, and in fact, necessary work. If you try to stamp out failure in R&D, the one thing that will certainly get stamped out is success.
3. Everything you have always learned about the 7 wastes applies in small measure in R&D. Yes, laboratories can be made far more efficient, especially production and testing laboratories. But in the end, these wastes pale in comparison to bad thinking. One great thought can transform an entire industry, but the wrong experiment will never generate that delight when a customer opens the box or looks at the price tag. Only good, creative thinking and experimentation will get you there.
Stamping out these wastes while preserving valuable learning is a wide subject in itself, so where can you start?
If you want to get rapid improvement in good thinking you can do two things:
a) Get it in writing. The easiest thing you can do is to ensure that every researcher predict in writing, the outcome s/he expects from each experiment. A pharmaceutical company applied this single measure, and cut the number of experiments needed to get a quality target pharmaceutical from ~30,000 to ~400. That is a LOT of improvement in a small box.
b) Insist on structured thinking. If you want to improve thinking further, and begin to solve the “re-learning” problem, introduce A3 thinking and ensure that every experiment is part of a larger A3. By ensuring that the larger problem is known, and that every experiment supports the larger problem, and every problem is solved using a structured (scientific method) approach, you ensure that unnecessary experiments are minimized, learning is maximized towards the project’s purpose, and the results are held in one place for future reference.
At one company I worked with, the research division lost fully 40% of its staff during a layoff. Most teams, lacked a structure or historic understanding of all problems required for success, and had to rebuild their work almost from scratch. By contrast, one team held their project in a set of linked A3s. During the chaos of the reorganization, that team was able to pick up key scientists cast adrift when other teams collapsed. With A3s providing a clear structure and understanding of prior work, newly engaged scientists could pick right up where others had left off. While other teams restarted from scratch or dissolved entirely, the A3 team hardly lost a step, regaining full operational momentum within days of the layoff.
4. Begin using trade-off curves and checklists. These two devices capture and deliver optimization knowledge with a glance. There is a great body of literature and engineering coursework on trade-off curves, and I urge you to begin there. And while you implement curves and checklists, fight hard to implement sharing and storage using old fashioned hard copy processes. If you can get learning and sharing process right using hard copy processes, chances are good that you will avoid the waste of useless software when/if you do apply IT tools.
5. Finally, if you want to learn where customer preferences lie, or better yet, shape customer preferences themselves, you need to fail often, fail fast and fail cheap. This means your experiments have to be at the lowest possible scale for meaningful learning. Batching of experiments (linking the same or multiple types of experiments together in one prototype or test cycle) leads to the same wastes and failure modes as batching in a plant. Cycle times increase, complexity rises, costs increase and failures get hidden until further on when costs are higher. Breaking “large scale thinking” takes some careful thought on its own. The challenge here is to create experiments that deliver only one piece of learning, and do so with the minimal amount of cash, time, people or materials. Anything larger, slower or more costly sinks unneeded cost into failure, and rarely builds additional success.
In the end, if you are lean practitioner in R&D, minimizing re-learning, maximizing the quality of thinking, and re-casting your view of failure can rapidly improve the quality and speed of your innovation.