Experimentation for Non-Scientists

As I get farther in my career, I find myself valuing the ability to experiment above almost everything else. Great teams sustain outstanding execution on their processes, but legendary teams challenge their processes and deliver improvements that dwarf what was previously possible.

Why is that? In rapidly changing industries, it’s not enough to just do the same thing in the best way possible. Customers change behaviors, tech creates new opportunities and employees get bored. So therefore, it often is more important to drive purposeful learning in addition to just executing again and again (thank you Lean Startup).

Learning happens when we experiment. We see what works or what doesn’t work. We then try again with that new information, continuously building on what we know. Sounds great, right?

Unfortunately whenever I mention experimentation in a corporate setting, I sense a lot of discomfort. I mean, how can I expect people to experiment without a laboratory? Feel free to add in many more science related jokes, as I’ll spare you from mine.

So this week, I’ll give a quick snapshot of how experimentation can work in a corporate setting and give some magic tips for making it stick.

Pre-Requisites

To experiment in your team, the key point to remember is that you really don’t need much to get started. In reality, you’ve probably been doing mini experiments for years now, but without celebrating successes or setting up a regular way to build upon your outcomes. Time to change that!

To set yourself up for success, I would suggest taking the time on the following before jumping in:

• Pick a goal that your team is struggling to achieve (because goals)
• Set up a place to log your ideas and track your experiments (Excel anyone?)
• Carve out a small portion of your day to focus on progressing your experiment (even just 15 minutes)
• Agree on a small timeslot in your team meetings to discuss experiment updates

That’s really all you need to get started. No fancy test tubes, beakers or Bunsen burners (sorry, I know I promised!).

Seriously though, that’s it. Things like fancy software to track or measure your results are nice but not required. Obviously, that would be helpful in the future but getting started is so much more important. The rest tends to sort itself out with the momentum of success.

You don’t even need sponsorship initially. As long as you are creative with your time and don’t impact your current responsibilities, this is something you can start on your own without depending on others.

The Process of Experimentation

Once you are ready with a goal and an idea to work on, it’s time to start experimenting.

The steps that follow are derived directly from the Scientific Method, but are pared down for us corporate junkies. Nevertheless, you are experimenting in the truest sense of the word while following this process.

Step 1: Measure your current state

Depending on what your goal is, you’ll need to get some sense of how you are doing now. This is important because you’ll want to take the same measurement at the end of the experiment to see if anything changed. Also, if your current state is very close to your end goal, you’ll need to adjust the goal to be more aspirational.

Check out Thomas’s tips on measurement and look to work with what’s available. For example, manually track the time to reach a client resolution if process data isn’t available. Go around and survey internal teams if satisfaction is difficult to determine. Do it a few times and use the average as your current state value.  

Step 2: Choose ONE aspect of the process to change

Once the current state value is known, take your idea and determine what should be changed in the process. Its very important to only change one aspect of the process at a time. Otherwise, you won’t know what impacted your current state value and what didn’t matter in the end.

An example would be if you are trying to reduce process time and your idea is to modify the procedure in a few ways. Pick a single step to adjust each experiment, see how that impacted the measurement and then move to the next step incrementally.

Step 3: Perform the process for a certain amount of time

Once the change has been made, perform the “new” process for some time. Depending on the volume of the process, this may be a few hours, days or even weeks. This can be a bit of an art, so just think of how many examples you need to really be sure you are getting an accurate measurement.

Step 4: Measure and Analyze

Once you feel you’ve processed enough, take the new data in the same way you measured the initial current state and calculate your new value. You’ll see that either you got closer to your goal, got farther from your goal or nothing was changed at all.

If you got closer to your goal, congratulations! Keep the change in your process and then look for the next idea to get even closer. If you got farther away, revert back to your previous process and try something different.

Follow the experimentation process again and again with new ideas until you are old enough to retire.

Tips for Best In Class Experimenting

As you can see, this isn’t rocket science. But the reality is that many individuals and teams struggle to implement even the most simple experimentation functions.

For some final thoughts, I have listed a few top tips to help ensure that experimenting becomes a deep rooted part of your team’s culture and something that everyone involved can enjoy.

• Celebrate each win. Make sure you are continually bringing the team updates and convey positivity when something works. Others are sure to follow
• Integrate experiment tracking into your production tracking. Having them side by side will make it feel more like an expected activity across the team
• Get your managers involved. After a few small successes, they will be likely to sponsor more of your time to deliver even bigger experiments
• Separate out larger ideas that may need other resources. They are great to have listed but focus on what you can act on initially

As always, we look forward to hearing about your own experiments in the near future.

Happy Experimenting!