by Peter Merrill
I wrote my first Innovation Imperative column for QP in January 2012,1 and my practice has always been to take the issue’s theme and address it from an innovation perspective.
The January 2012 issue’s theme was risk. I discussed how risk enters the innovation process. I explained how risk is internal and external, and how organizations have become extremely risk averse as we have increased our understanding of risk. However, there is evidence that the more radical or innovative our new offering, the more profitable the outcome. But, the more radical the offering, the greater the risk attached to it. In short, innovation is a risky business.
In this month’s column, I am going to move beyond identifying the nature of the risks we must address and look at how we manage those risks.
My first key point is one you have probably heard before—fail early. The innovation process has two phases: creative and execution. The budget for the creative phase is typically only 20% of the total cost of getting a new offering to market. Clearly, the more we can weed out the weak offerings in the creative phase, the better the outcome.
The creative phase has two steps:
- Identifying the opportunity.
- Connecting the conceptual solution.
Focusing on the second step, physicist Linus Pauling said the best way to get a good idea is to get lots of ideas. To generate lots of ideas—and especially to generate radical ideas—we must allow these two steps in the creative phase to operate in a loose or free mode. At that stage, in fact, we avoid even thinking about risk.
But after we have generated our potential solutions, we do need to address risk. To do that, we need data. Unfortunately, creative folks do not always enjoy data collection, but it is vital to manage risks.
Some of the primary areas in which we need data (which, by the way, likely will not be crisp) are as follows:
- The time needed to develop a working solution. At this stage, we are only at concept.
- The probability of being able to produce a working solution.
- The cost to produce a working solution.
- How radical is the new solution compared to current solution?
- How easily can this solution be copied by the competition?
- Do we have in-house competencies to produce the solution?
- Who are the new suppliers we will need, and what is the risk attached to those suppliers?
- What are the delivery chain choices and the attendant risk?
- What price will the market bear?
- The likely return on investment (ROI) of the new product or service.
Those of you who live in the R&D world will be familiar with the necessary data for developing a new product. Much of this data will have wide tolerances attached, but that does not excuse us from collecting the data. It can often be helpful to provide support for data collection to the folks working on the proof of concept if data collection is not their strength.
Another issue to be aware of is the introduction of bias into the data collection. This can happen so easily and is difficult to detect. Somebody sees a wild idea and says to themselves: “Oh, I’ll kill that right away,” meaning, the person will use data to reject the idea. And yet, that wild idea may be the perfect solution to the opportunity we are addressing. Us engineers tend to be good at data collection and analysis, but we are not good at accepting change and are extremely risk averse.
We may also have to generate data by modeling or piloting a solution. Software guru George E.P. Box said, “All models are wrong, but some are useful.” Never lose sight of that, but we must still do everything we can to develop data that will guide us in our choice of solutions.
In my early years as a chemical engineer, I worked on some exciting pilot plant projects. They really were great fun. I worked on, among others, desalination by vapor recompression, radical countercurrent liquid-liquid extraction and the early days of carbon fiber.
We have to be careful to ensure these projects do not take on a life of their own. They are there to generate data. At the same time, we must not isolate the people running a pilot. They must be connected to the bigger project.
After we have the data—and unfortunately it will not all arrive in perfect form—we have to start making decisions. The body of data is probably going to be complex, but we must not lose sight of our end goal. We want the most radical new offering with the best ROI at the lowest risk. Call it utopia if you will, but that is where we aim. This idea can be illustrated in a three-dimensional risk tool (Figure 1).
Lead the crowd
British financier James Goldsmith said, “If you see a bandwagon, it’s too late.” In other words, don’t follow the crowd. It’s far more profitable to lead the crowd.
One of the tools I like for evaluating risk data is a modified version of design failure mode and effects analysis. There are a number of other tools and combinations of tools, but I don’t want to dwell on the use of these tools in evaluating risk levels. I am going to stay focused on managing risk.
After we have collected the data, we need to make decisions. It is essential to narrow focus to manage risk. The decisions are based on business need and capability, and form the critical forward strategy of any organization. Without new offerings, all businesses will eventually die.
Unfortunately, in a science and technology-based business, the link is not always healthy between the research people producing data or proof of concept and the strategy people making the decisions. Ironically, the less scientific the organizations, often the better the link.
Another factor is organization size. When I worked in R&D, I was in a major chemical corporation and the research division was almost an entity of its own. Connecting to the forward strategy of the business was a challenge.
With these factors in play, strategic decisions often become minor adjustments of last year’s activities instead of forward-looking strategies with innovative new offerings.
The job of strategists is to make choices and develop the new product portfolio. The portfolio must be a mix of short, medium and long-term projects. The size of the mix will depend on the size of the business. A small to medium-sized enterprise will have probably only two or three selections. However, we should never put all our eggs in one basket and focus on one new offering—that is far too high a risk. Figure 2 (p. 43) shows how project risk relates to project time scale.
The packaged goods industry invests heavily in short-term offerings, putting more than 30,000 products a year on the shelves. More than 70% of those products die within a year. If this is low cost, then that is a reasonable strategy. But every business must have a long-term strategy in which it looks as far as 10 years into the future. This is the “seed corn” for tomorrow’s business.
For example, Xerox found that it had to wait an average of 7.5 years for an acceptable ROI on its best innovations. Unfortunately, the pressure for short-term performance inhibits most companies from taking this long-term view. But it’s important to only use ROI as a guideline. Think logarithmically: Is revenue going to be $10,000, $100,000 or $1 million?
The development stage
After decisions have been made and the portfolio agreed upon, we are still only at the start of managing risk. There is often the false assumption that risk will not change. But we now leave the research stage and enter the development stage.
Organizations are often structured incorrectly by bundling R&D together. Research should partner with marketing; development should partner with operations. Not many organizations do this, but for successful innovation you need to give it serious thought.
In the development stage, we take a concept and develop it into a working solution. The development stage continues to gain knowledge about the new offering. As a result, the risk numbers will change and we must manage this.
One of the toughest challenges here is killing an idea that does not pan out but that passed the first cut with the strategic planners. This idea will almost certainly be somebody’s “baby” and there will be a lot of emotion involved. People will even mask new data to keep an idea alive.
New data must be thoroughly analyzed at each design review meeting. Design review meetings are notorious for not clearly recording decisions. The design review also must be hardwired to the strategic planning activity. The skills of the strategic decision maker lie in allowing enough rope in development, but not too much for people to hang themselves. It’s also important to remember that this refreshment of risk data is not an annual activity. It should be done continuously and should typically be reviewed on a quarterly basis.
I have focused on risk attached to the new offering, but there is another important area of risk that is often neglected: external risk. This is split into supplier or subcontractor risk and delivery chain risk. In my January 2012 column,2 I quoted data showing that with four new suppliers, three had a 90% probability of delivering and the fourth had a 40% probability. The overall probability of the new offering coming together is only 29%.
We must evaluate and then manage supplier risk. It may even mean purchasing a supplier, but that is often bad strategy. There are many ways to mitigate supplier risk, including the use of inventory, on-site presence and technical support.
Delivery chain risk is often more difficult to manage. This is where we start to give away our secrets. A good strategy here is to partner in the early days with a chosen customer or distributor. I used this strategy frequently when I ran a leading brand in the United Kingdom. It is an excellent way of improving speed to market.
Risk in innovation
Initially, we create many options in the creative phase while risk is low. Success depends on collecting good data while still in the creative phase of the new offering. We then narrow our focus in creating the new product portfolio based on our aim to produce the most radical new offering with the best ROI at the lowest risk.
To achieve this, we must retain a strong link between the subsequent development stage and the strategic decision making as risk data becomes more precise.
- Peter Merrill, “Risky Business,” Quality Progress, January 2012, pp. 50-51.