Developing a culture of experimentation on your team

Most companies support the idea of incorporating innovation into their business strategies, as it can help increase market share and generate additional profits through new products or service offerings.

But too often, these same companies fail to realize that innovation doesn't occur without experimentation. To successfully innovate, you need to conduct experiments—lots of experiments. Some will succeed, and lots more will fail. By definition, experimentation is "the process of performing a scientific procedure, especially in a laboratory, to determine something" as well as "the action or process of trying out new ideas, methods, or activities." Innovative organizations don't isolate such innovation practices to certain segments of the business; they cultivate an attitude of experimentation throughout, weaving it into the very fabric of the entire organization.

So the question for your innovative organization becomes "How do we create a culture that allows us to be comfortable with trying out new ideas, methods, and activities using a scientific procedure?" Having a healthy culture of experimentation is the only option if teams wish to innovate. And organizational leaders can play a significant role in fostering that kind of culture.

Before teams can even begin operating in an experimenting mode, leaders must shift their mindset to operate not like a know-it-all, but as if the world is truly a complex space, that we don't know it all, and that we can't know something unless we try it out.

When I visit companies and hear leaders say things like the following, I quietly take them around back for coaching:

• "Your demo failed." (This is usually accompanied by a scolding face.)
• "You need to run everything by me before conducting a test with the customer." Screeching brakes are heard as innovation comes grinding to a halt.
• "Let's do a dress rehearsal showing the customer how this operates before giving them the feature to manipulate." No, no, no—put the software in your customer's hands as early as possible. You'll learn what's wrong with your user interface. You'll reduce risk, increase quality, and build a relationship with the customers.
• "Well, obviously they need [insert feature X that will get me my bonus if released within the year]." Carefully think about how your compensation program may hinder creating an experimentation culture.

Taking a deeper look at the first three points, fear of failure is overwhelmingly the operating model. The last point is a desire for monetary success over doing what's right for the customer, regardless of where the idea comes from.

Instead, we want leaders involved in conversations to encourage experimentation. Those conversations sound like this:

• "What is the customer's problem? Have you observed this?"
• "What is your hypothesis?"
• "What are your critical assumptions that must be true for your idea to work?"
• "Do you need help designing an experiment to test if your hypothesis is true or false?"
• "What can you test to (in)validate your hypothesis?"

So how can you move from the first kind of conversation to the second? Here's my advice: Don't wait for the perfect moment—just start!

Here are 9 pointers to help get you start experimenting as a team:

1. Don't jump into the solution space. First, define your problem. State it as a hypothesis.
2. List all of your assumptions.
3. On a 2x2 matrix, rank each assumption in terms of uncertainty versus risk. Identify the highest-risk, most uncertain assumptions before moving on to step 4.
4. Create a simple test experiment that you can begin to work on today. Action is key; however, think low-fidelity, rapid prototyping to be able to run this test. (For ideas on how to start prototyping, try the exercise at the end of this article.) In other words, be able to collect as much information with as little effort as possible. The key is to run your experiment with real people to get real results. The free test cards from Strategyzer are wonderful to help get teams thinking with a test mindset. Each card begins by stating the hypothesis, then the test, an accompanied metric, and criteria for success.
5. Gather the data and record everything: The data you collect and record will guide you further.
6. Review results as a team. If you use the test card, you will have thought about criteria for judging whether your test was a good one. Ask questions! What did you learn? Do you need to change your hypothesis based on what you learned? Do you need to do a new experiment? Do you need more data?
7. Share the results with the rest of the organization—especially the failed experiments.
8. Celebrate the learnings. You don't need to copy Spotify's Failure Wall (described in Henrick Kneiberg's Engineering Culture, Part II), but the point is to celebrate what you learned, and how you want to conduct the next experiment to incorporate what you want to change.
9. Rinse and repeat!

Steps 7-9 are crucial for leaders to help make this culture change stick. Leaders must hold the space where anyone can experiment. Anyone can run a failed experiment because these failed experiments mean we learn. You need to make it safe to take risks and to remove impediments. Start today with your own new mentality that will sweep across the entire organization: Instead of "no failure is allowed here," adopt the mindset of "We won't know until we run an experiment." It's your only option if you wish for innovation to happen.

A brief example: Dollar Shave Club

Dollar Shave Club's hypothesis was that men don't need fancy razors with lots of fancy features, nor do they want to spend time shopping for them. Their experiment involved creating a 1-minute video sharing the problem and offering a solution by inviting viewers to link to a landing page where they could place an order. Within 48 hours, they knew the experiment worked: They had received more than 12,000 orders.

[See our related story, Introducing the Open Organization Workbook.]