Successful innovation always requires experimentation. For example, in previous podcasts I’ve talked about rapid prototyping.
The key points in this podcast are drawn from a Harvard Business Review article titled Enlightened Experimentation: the New Imperative for Innovation. The author draws his information from research in the pharmaceutical, automotive, and software industries. The focus is on for rules learned from that research.
The first he calls organize for rapid experimentation. He says, “the ability to experiment quickly is integral to innovation: as developers conceive of a multitude of diverse ideas, experiments can provide the rapid feedback necessary to shape those ideas by reinforcing, modifying, or complementing existing knowledge.”
The author gives the example of the car manufacturer BMW. In efforts to improve the crashworthiness of their cars, they historically would build expensive prototypes. These were not only expensive but often took months to build. In the course of the overall development cycle for a new model, the time to test with expensive prototypes typically returned crash results too late to effectively integrate all the learnings into the new model development.
Today the company uses virtual experiments – this is where crashes are simulated using powerful computers. This has significantly reduced experimentation costs and dramatically speeded up the receipt of crash results information. This allows for rapid experimentation and learning. Ultimately the virtual experiments are confirmed in actual prototype testing, but by that time the probability that the virtual experiments are confirmed in actual testing is very high. By utilizing this process, BMW significantly improved the side-impact crash safety of its vehicles. If they used the old methods, the physical prototypes would cost about $300,000 which was more than the cost of all the 91 virtual crashes combined. Yes, that’s 91 virtual crash experiments.
Use of small groups also helps rapid experimentation. You just need to make sure that you have the right people in a small group. For example in BMW’s case, they had designers, test engineers and manufacturing engineers as a part of the team. Another benefit that helped speed up experimentation for BMW was that several tests could be conducted simultaneously by running them in parallel.
The authors second rule is fail early and often, but avoid mistakes. The author says, “Experimenting with many diverse – sometimes seemingly absurd – ideas is crucial to innovation. When a novel concept fails in an experiment, the failure can expose important gaps in knowledge. Such experiments are particularly desirable when they are performed early on so that unfavorable options can be eliminated quickly and people can refocus their efforts on more promising alternatives.”
Some people are uncomfortable with a rule like this because it involves failure. I’ve always looked at failure as a teachable experience. If I am able to convert what I’ve learned about a failure into a future success then failure is a very good friend indeed.
A company mentioned in a couple of previous podcasts, IDEO, also highly values the ability to have rapid experimentation early on in an idea’s life to facilitate rapid learning from both what worked and did not work. Subsequent experiments incorporate learnings which typically improve results, but there still may be shortfalls that need to be learned from. The cycle continues by then incorporating those learnings into the next version.
There are some important basics when it comes to experimentation. There needs to be a clear objective that outlines what you anticipate learning. There needs to be a hypothesis detailing what you expect will happen from the experimentation. Experiments also need to control the variables. If a single experiment includes multiple changes and performance improves dramatically, you need to know how much each change contributed to the performance improvement.
The author distinguishes between failures and mistakes when he says, “mistakes produced little new or useful information and are therefore without value. A poorly planned or badly conducted experiment, for instance, might result in ambiguous data, forcing researchers to repeat the experiment. Another common mistake is repeating a prior failure or being unable to learn from that experience.”
The authors third rule is anticipate and exploit early information. Deep and broad learning about an innovative idea early on is critical to success. For example, research in the software industry suggests that late stage development problems are 100 times more costly than mistakes caught at an early stage.
Not only are late stage problems much more costly financially, they can also cost a project precious time in getting to market. Speed to market can be a major benefit. For example, a pharmaceutical• product that can reduce its development time by six months has the benefit of effectively extending patent protection by six months. For products that might have a short life cycle, reducing development time by six months can extend a trend based lifecycle and substantially improve profits.
Consider the before-and-after example for Chrysler. In 1993 when they developed the Concorde and Dodge Intrepid models they used a physical process for fitting the powertrain and other components into a prototype automobile. This process took three weeks and many attempts before they achieved success. By 1998 they were using digital mockups powered by computers. Instead of taking weeks to accomplish the task, it was completed in 15 minutes.
The author’s fourth rule is to combine new and traditional technologies. It is not unusual for companies to become enamored with new technology and view the new technology as a replacement for their traditional or historical technologies. There can be significant risks in thinking this way.
New technologies can be buggy early on, which reduces their effectiveness below the previous technology. The new technology may not perform all of the capabilities of the previous technology.
As is often the case with technologies, it is highly desirable to have an integration of multiple technologies. When this is done effectively, you can experience the benefits of both old and new technologies.
In the pharmaceutical business, new technologies have significantly increased efficiency and speed when developing new chemical compounds. Researchers no longer need to carefully develop one compound at a time. Instead they can use what is called combinatorial chemistry. This allows chemists the opportunity to develop numerous variations in parallel built around similar hypotheses or building blocks.
There is considerable information here that has varying applicability to help your business. I want to share my thoughts on how you can best use the information shared in this podcast to help you sell more and make more.
Probably the most universal of the rules in this article is the one about organizing yourself for rapid experimentation. Regardless of the business that you are in, developing prototypes ranging from fully operational to looks like prototypes is a proven rapid learning technique. Even prior to this step, producing multiple written concepts and variations on those concepts can be very valuable when you share them with potential customers either in a qualitative or quantitative research environment. Typically the best learning cycle is from concept – that is, a written expression of the idea maybe with a rough visual or two – to the development of various types of prototypes that are shared with potential customers in either a qualitative or quantitative research environment.
The second rule can also significantly benefit your innovation program. The benefits of rapid experimentation and failing early and often can really be seen as one rule. It is very, very important that you not shortchange the upfront learning program on a new idea. Just because you get positive feedback at an early stage, does not mean that you solved the puzzle to marketplace success. You need to learn from multiple variations of the idea. Upfront aggressive learning programs will absolutely increase your chances for eventual marketplace success – or if one of the learnings is it’s not a good idea, you can save an expensive failure.
I hope these insights help your innovation program so that you can sell more and make more.
