Another Seth (Godin) posts on his blog:
Wait, I was confused. There’s a sure-fire recipe for delicious chocolate chip cookies. There is in fact a magic formula.
For businesses, not so much. There isn’t one secret, one process, one solution. Instead, there are a thousand or maybe a million.
It’s not a jigsaw puzzle, it’s a strand of DNA, easily rearranged and sometimes it even works. For a while.
First, recipes for delicious chocolate chip cookies are maybe as numerous as recipes for successful businesses.
Second, I agree with the strand of DNA analogy. But, I like to think of it as a trial-and-error experiment. That’s what different configurations of DNA are as well – trial and error experiments.
Every new product, every new business, nearly every decision made in the running of the business is a trial-and-error experiment with an uncertain outcome.
The business experiment tests whether customers find enough value in whatever is being sold to trade what’s needed to encourage the producers to make more. Some experiments work out and many others don’t. There’s no sure fire way to tell if you’re on to something without trying it.
While I agree with Godin that there is no sure-fire recipe business success, I do believe there’s a way to improve chances of finding successful businesses with good experimentation.
For existing businesses the lesson is to try many new things on small and inexpensive scales. Many businesses flip-flop this. They try a limited number of expensive experiments.
Mature companies forget they were founded as a result of a trial-and-error experiment. Some never even realize it.
Managers appoint themselves as the judge of what will work and what won’t and they filter out many ideas based on their own preferences, never realizing that their preferences aren’t worth a heck of a lot.
Mature companies become old and clunky because they reduce experimentation, sometimes without realizing it. Even after choking off conscious experimentation, most companies have natural experimentation happening within their business and they tend to ignore the results or kill them if the new process or product deviates too far from their comfort zone.
I remember sitting in a meeting discussing a new experiment at one company. The 20 people in the room from different parts of the business were each interested in getting their stamp on the experiment so they could appear to be adding value and have something to plug in their next performance appraisal. None of them had been a part of any previous successful experiment. The experiment was a Homer (as in the car designed by Homer Simpson). It had a little something for everyone and worked for no one.
Folks from this company will tell you that they experiment. They just don’t realize that they experiment poorly. They don’t conduct enough experiments, don’t have the right people working on them and the experiments they do conduct are subject to too many political stakes.
Here are my suggestions for companies who are interested in tapping into the power of experimentation:
- Increase the number of experiments.
- Make the experiments as cheap as possible.
- Measure success on actual customer response and feedback.
- If you need a statistician to tell you if the experiment worked, it didn’t work or it didn’t work well enough. You want experiments that work really well.
- Keep an eye out for natural experiments. These may take the shape of an accident that has a good result or someone in a remote location doing something a little different that’s generating good results.
- Associate successful experiments with people. Don’t expect anyone to bat a 1,000, but notice and reward that some people have a talent for hitting on successful experiments. You want to keep these people around and let them have some runway and reward them when they do hit on something.
- Resist temptation to impose arbitrary preferences on experiments. I’ve seen leaders kill any hopes of a successful experiment by imposing their own preferences into it (e.g. “it must use our brand name” or “it can’t cannibalize our existing business”).
- Kill failed experiments. I’ve seen failed experiments carry on far too long because the experiment happened to be the pet project of someone with political power in the organization…
- …but don’t be afraid to evolve experiments. Sometimes an experiment needs to evolve a little before it hits.
- Start small and expand slowly. Another crazy thing big companies do is roll something out big that has little or no experience, battle testing and evolution. These are either ego-driven projects, hail mary acts of desperation or just bad business management. These types of projects may have been the origin of the phrase epic fail. Management tends to over react to gain a first-mover advantage, but that’s a myth. The eventual winner tends to be the product that is most battle tested and evolved to best suit the needs and preferences of the customers (e.g. Google vs. Yahoo!, MySpace vs. Facebook, Lotus 1-2-3 vs. Excel).
- Use an entrepreneurial model with small project teams and get those teams away from HQ so they won’t be influenced by office politics.
Show me a company that is lagging, let me speak to a few people there and I can likely tell you how they are snuffing out the power of experiments. Show me a successful company and I can find out how they came upon the successful experiments and see if there’s trouble in the future due to a bad experimentation model.
To use a band analogy, some companies end up being one-hit wonders while others endure. And for the very same reasons. One hit wonders may make enough money and decide they no longer need to experiment with new music. Enduring bands never seem to lose that experimental motivation. They continue to experiment with music and continue to hit upon new hit songs.
To sum up, the secret to good business is good experimentation.