Beth was nervous. She was six months into the most demanding leadership challenge of her career — building a high-growth-potential (but risky!) new business inside of the large company where she had worked for over a decade. It was becoming clear that the business was not progressing as planned. It was behind schedule. It was over budget. And its prospects remained murky.
Beth was nervous because at an important meeting the next week she would have some explaining to do. In her company, the rule was simple. If you deliver on plan, you are rewarded.
To some extent, Beth felt that this was unfair. The business she was managing was uncertain. Nobody could predict its future. And to some extent, Beth’s superiors understood this, too. So at the meeting, when Beth asked that her business be treated with an “experiment and learn” approach, they agreed.
Unfortunately, Beth had really just won a stay of execution.
The notion of experimenting and learning made everyone feel good for the moment. It had a nice ring to it. It sounded like something that you have to do in order to innovate. But neither Beth nor her superiors really understood what it meant.
Actually, experimentation is easy. Experimentation is simply means trying something new and unproven — and acknowledging that you cannot reliably predict the outcome. But learning, what was that? How do you know it is happening? Why is it important?
There are some business contexts in which the notion of learning is straightforward. If you want to know if you are learning to manufacture a new product, measure the trend in defect rates or unit production costs. If you want to know how quickly you are learning to sell, measure the trend in cost per sale, or success rate at turning qualified leads into closed deals.
In the context of testing an experimental business model, however, the meaning of learning is less intuitive. It turns out that a very specific kind of learning matters most. Beth needed to learn to predict. She had not made good predictions in the first six months — but forecasts are always wild guesses when a business is brand new. The question was, could Beth make better forecasts now? If so, she was learning.
Learning can sound like nothing more than a “feel good” objective — distant from bottom-line business imperatives. But learning to make better predictions is deadly serious. It is linked directly to the bottom line. If you can predict, you understand which actions lead to positive outcomes, and which do not. You can make a sound judgment about whether a business is a winner or a loser. When you learn to predict as quickly as possible, you minimize time to profitability. You minimize risk. You minimize capital consumption. And you maximize the probability of success.
Learning to predict, however, is not something that just happens. Lessons learned do not magically fill the minds of those who come to work with an “experiment and learn” attitude. Learning to predict is a process. It requires analytical rigor and discipline. The key step is carefully resolving and explaining differences between predictions and outcomes.
As simple as this sounds in concept, many things go wrong. Predictions should be treated as though they are sacred. They should be carefully retained, along with all of the underlying logic, so that they can be reexamined later.
Instead, predictions are abused. Sometimes they are casually discarded. After all, initial plans are based on a tremendous amount of guesswork. It can seem embarrassing to go back and analyze a plan that was so wrong.
Or, predictions become rigid. This can happen as senior executives insist on maintaining a strong culture of accountability. Or, when the leader of a new business views any revision in goals as backing away from a bold future. But learning is about making predictions better — not faithfully adhering to original goals.
Finally, predictions are manipulated. Perceptions of performance are important game pieces in the ongoing battle for power and influence within an organization. Managers may be able to achieve a political win by altering expectations. But doing so disrupts the learning process.
Beth may have won the battle. If she is to win the war, she will have to do more than delay the day of accountability by preaching “experiment and learn.” She will have to focus on improving predictions. And if her superiors want to help, they will have to be more than just judges. They will have to join Beth in the learning effort.