I experiment a lot. I attempt something, measure my results, use those measurements to see what’s going on, reassess and change my attempt for the next time. I do this with my weight training at the gym, my writing, my book writing and production, and my work. When I experiment, I can inspect and adapt. I can refine. I can improve.
I did this when I was a developer. I noticed the kinds defects I wrote. I kept a log. What kinds of infinite loops did I write? I used a log to keep a list. That is a kind of measurement. As a writer, I notice what my copyeditors notice. I keep a list. As a speaker, what kinds of feedback do I receive? I keep a list. These are measurements. I can experiment with my work, get the feedback, and try something else.
As a writer, I experiment “internally”, too. I’ve experimented with writing fast and writing slow. I discovered that writing fast, getting the words out without editing is faster and better for me. I do a better job at getting my ideas across to my readers when I write in Markdown, just getting the words out. When I write in a text editor, I am slower, and I tend to edit as I go. My results are worse.
At the gym, I often think, “I can’t.” That’s my default position. (I keep saying I’m a work in progress :-) But I have learned that I probably can. I have learned to think of the first set as an experiment. “If I experiment with this first set and see what happens, I can use these results to inform my second and third sets.” That’s also the growth mindset.
But notice, I said experiment, not try. I said I measured, and used the results of those measurements to reassess and change, based on the results of my measurements. That’s how I know I’m experimenting.
If I was “trying,” instead of “experimenting,” I wouldn’t be measuring. I wouldn’t be adapting based on my measurements.
When we solve problems, sometimes we try—an effort-based approach. Sometimes we experiment—a measurement-based approach. I find it useful to discriminate between the two.
Trying is an attempt without measurement behind it. Experimenting adds the idea of measurement. We will inspect, measure, assess what we have completed. We might even adapt what we do, before we abandon it.
If you’re eating, trying is fine. You don’t have to experiment when you eat. But to learn, to change, to grow? You need to experiment. How else do you know what works, as opposed to what you think works?
This week, I’m at Agile 2014, the big conference about agile. Some people are confused about whether they are trying or experimenting. They call what they do experimenting, but they aren’t measuring anything. They aren’t using measurements to assess and change, based on their results. They try something, declare it a success or failure, and continue. But they have no data.
The problem is, you can’t inspect and adapt without knowing your data. Well, you can. That’s called randomness. You can do anything you want. But if you want to experiment, based on reality, you need data.
Trying new things is great. Experimenting is great. Let’s not get the two of them confused. When you try, you make an attempt. You don’t necessarily have any data to back you up. With experiments, you have a hypothesis, you collect data will explain your hypothesis. Let me rephrase Yoda, “There is no try; there is only experiment.” Okay, my geekiness is showing. Experimenting is a form of doing.
I met someone briefly yesterday who said, “Our developers always lowball their story estimates for an iteration. They can’t break their tasks down enough. How can we help them to not do that?” (For my non-software readers, stories are the requirements, and the iteration is a one- or two-week timebox.)
I asked, “Did they talk about this in the retrospective? Because this is not an estimation problem. Well, it might be. But it’s almost always a story-is-too-big problem. I would start by looking at the stories, and not by breaking down the tasks. What experiments have you tried?”
She looked at me with that Oooohhhh look. You know that look, the one where someone says something to you that you know you should have thought of yourself, but you didn’t. One of the big transition problems in agile is that teams have trouble making small stories. We know this. Teams need to experiment with their right way to break the stories down.
Try is one way. Without data, you cannot win. Experiment is better. Measure and improve. Inspect and adapt.
So, gentle readers, the question of the week is: Are you trying or experimenting? If you know the difference, your adaptable problem solving could improve.