Having an experimental mindset helps leaders and teams open a variety of possibilities and a free flow of ideas. Seeing the things you try as experiments orients you to collect data and evaluate whether the experiments or things you try are working or not. Sometimes a team will luck into things that work but many times they don’t. When things don’t go as hoped, we fail. An old friend of mine says, “Education is what you get when you didn’t get what you wanted.” And education can feel like a consolation prize in the race to innovate. It isn’t about whether it feels bad to fail (it does), but how to emerge resilient and ready to experiment again. To fail better, experiments need to be safe-to fail, small enough to iterate often, and there needs to be a plan in place for data collection and decisions about next steps.
First, experiments need to be safe-to-fail. A safe-to-fail experiment is a small scale experiment that will not negatively effect the business but big enough to notice an effect. One strategy that can help make experiments safe is to make single variable experiments whenever possible so you can stop the experiment quickly without being stuck if the outcome isn’t favorably. A single variable experiment means you only change one thing. Large implementation of new software is where I often see organization’s experimenting in ways that aren’t safe-to-fail. It is worth thinking about how to try it first in a safe-to-fail or in a single area of the business. In 2015 Target closed all 133 stores in Canada after $2 billion in losses. They had supply chain issues, problems reconciling metric measurements to fit things onto shelves, problems with online shopping, and a whole myriad of challenges that could have been solved one at a time if they had opened just one store as a prototype to solve the challenges of coming into a new market. For them, this would have been a safe-to-fail experiment. Instead they put their whole business at risk. Choose a small experiment to start with—it is more likely to be safe-to-fail.
Second, you should cycle through experiments until you don’t see a way to improve. This may feel foreign but after years of experimenting, it is clear when it is time to stop experimenting. This happens after a series of experiments and next steps leaves you with a satisfactory status quo. Try to experiment in a way that you can start right away. Experimenting shouldn’t take months of planning or a big budget. Figure out how you can start today. Use a prototype you cannot make yourself and see what you can learn before tomorrow. Then, alter and experiment again. The more experiments you try, the more work you are doing and the more your product (or service, policy, etc) will improve. Frequent iteration is the magic in experimenting and is the key to becoming a great experimental leader.
Finally, good data collection for your experiments will help you become a data-driven leader. Always ask, “What did you learn” and “What is the next thing you will try?” So many implementations are made of next steps, new software, new policy, etc without collecting any data on whether they work better than the old. Ask for data. Ask your teams to experiment and bring the data back when you are making decisions. It isn’t always possible to experiment, but do it anytime you can. It will safe you money, time and heartache when you experiment first. Collecting data will help you become a data-driven leader who makes sure solid decisions.
Your experiments need to be small enough to be safe to fail, you need to iterate often and you need a habit of good data collection to be a great experimental leader. These practices will allow you to feel more solid as a leader and you will feel like you are making decisions from solid ground. Your teams will be incrementally improving everywhere and your organization will be practicing intentional creativity and innovation in reliable and repeatable ways. What can you do today to improve as a leader tomorrow?