Experiments Aren’t Just for Scientists, Use This Proven Tool to Improve Your Offerings
Any business that is looking to launch new offerings, improve processes and solutions, or simply grow should be running experiments. There is an effective way to conduct experiments, and there are many not-so-effective ways. Throughout my career, I have been on both sides of that fence—once running hundreds of experiments in an attempt to prove that the product my team had created was right when other experiments were telling us it was not.
When you run experiments based on solutions, you have a tendency to think that if the experiment doesn’t produce the results you were hoping for, it’s a failure. That is a fallacy. There is no such thing as a failed experiment, only failed outcomes or failed hypotheses. Experiments cannot fail; they only teach you what to do the next time. Every experiment is an opportunity to learn and to design better experiments.
Experiments are a chance for incremental—or sometimes acute—learning. They give you the insight to make your product or service better. However, the truth that an experiment reveals can often be hard to swallow if ego is tied to the results. This, I’m convinced, is why more companies don’t adopt a culture of experimentation and learning.
With every new experiment you run, you get closer to a better offering, whether you are improving one that is already existing or developing a minimum viable business opportunity. The more experiments you do, the more you will develop increased knowledge, experience, and validation. Experiments are a chance to learn more about the feasibility of your solution and adjust it according to the results. The experiment itself is not your solution. It is what you learn through the experiment that leads you closer to your final solution.
There’s no way around it: if you want to develop an offering that is sustainable, you must run experiments. Chances are you have already run experiments by default to get where you are today. I say “by default” because much of what you do in business, including interactions with your customers, are essentially experiments. But when experiments are run without parameters, they take longer than they should, become costly and complicated, and don’t yield results. You may even miss the lessons that are offered if you’re not looking for them, and be doomed to repeat the default experiment until you do.
The trick is learning to run many experiments quickly. You can run a two-day experiment, a week-long experiment, a month-long experiment. It depends on the data you need to gather. But always run experiments with a clear purpose.
Don’t confuse a clear purpose with a perfect experiment. I’ve seen clients stall on running experiments because they want to design the perfect experiment. They want the experiment to answer all of their questions at once, so they begin to design it with many bells and whistles. One of the factors that discourages running experiments is making them too complex. Effective experiments can actually be quite simple. In fact, the combination of simple and fast often results in the best experiments. A sense of urgency is important when you are trying to accelerate the learning process. Remember that experimentation is not a one-time process. You are not trying to find your final solution the first time out. You are instead trying to learn as much as you can as you work toward the right solution.
An effective experiment includes these four elements:
Clear Purpose. An experiment without a clear purpose can easily lose direction. Simply completing the statement: I want to learn XXX so I can XXX helps you determine the intent of your experiment.
Clear Measurements. Determine what you will measure beforehand, keeping in mind that measurements may be quantitative or qualitative.
Clear Decision Criteria. After you outline the measurements you will collect, give yourself parameters. Determine the next decision you will make if your measurements are within a certain range. This allows you to have confidence in taking the next step—and knowing what that step will be.
Actionable Plan. Every experiment should have a plan that includes who will be involved, when it will start and end, where it will be conducted, and who will review the data collected. Without an actionable plan, it is easy to miss something or avoid taking action altogether.
To begin to conceptualize experiments, take a project and break it down into simple aspects you want to learn, to both reduce risk and improve your offering. Another easy way to design an experiment is to identify 3-5 major questions you want to be answered and convert those questions into experiments. Start small, and start with practicality. Consider an experiment you could run this week. That is how you begin to build your experimenting muscle.
As you begin to design and execute experiments, you will find that you become much more experiment-minded and experiment-driven throughout the business. You will understand that you don’t have to go to market to prove solution viability. You can start engaging with your customers right now through small experiments before you launch your solution widely.
