Innovation has been a longstanding buzzword in Silicon Valley, and with the rise of concepts like design thinking and circular design from IDEO, and Sprint from Jake Knapp and Google Ventures, many companies have begun the process of integrating innovation as a continual process (as opposed to goal) into their central work processes.
The next natural outgrowth after a few years of hacking away at innovation has become experimentation programs. What may look like the latest business management fad to some, to me looks a lot like the next step in business evolution. For any business dealing with significant scale, it becomes quickly apparent that the experimentation aspect of innovation requires a program all it's own. If you're just now hearing about this, a quick search for "experimentation" in " Jobs" on LinkedIn yielded a wealth of results, many from top companies. It is the next wave of this change.
Great! ... Kinda. The problem with this, like most business fads, is that while it's great so many companies are diving head first into innovation and experimentation, what's being lost in the sauce is often the rigor. I often advise well-meaning business leaders and teams that are devising experiments to run in their business (great 👍) but they aren't taking the relatively small amount of time it takes to put structure to these experiments. As I've discussed previously, innovation and experimentation can be very costly, so it' s important that we are taking intelligent risks, or it will quickly become a financial liability. Many companies (not naming any names) have started innovation programs only to scrap them a few years later. This has mostly to do with the how expensive it can be to run one—unless you're doing it right.
So do yourself a favor (and everyone else in the industry) and devise a structured procedure to test a well-thought-out hypothesis. Make sure you are delineating a clear and repeatable step by step process. Be explicit about your plan to elicit collect and analyze data and feedback, and whether or not you will be running multiple trials or running the experiment multiple times to understand your results. And for all our sake's please have a plan to communicate your lessons learned before you start. Here's a quick checklist I run for through for myself when I create a procedure.
- Is it repeatable? Could someone else understand these simple steps to repeat the process without having to give them a ton of background?
- Is the process testing the right things? Does it test a hypothesis by the same variables I gathered baseline measurements for?
- Is it rooted in a genuine problem statement, or observation about our business performance?
- Is it unbiased? (No, it's not.) But have I gone to reasonable lengths to eliminate bias from the process? (See here for a decent place to start.)
- Do I have a clear and explicit plan for how to review, organize and communicate my findings?
Once you develop a process, be thorough, systematic, and consistent about following it. Be down right religious. Use it to carefully collect new data and judge whether your procedure is effective at testing for the variables you are tracking.
It's not good enough to just try something new and call it experimentation. When we do things like this, we mistakenly give people the idea that experimentation itself isn't worth the investment, when in reality it's only not worth it if we aren't making our learnings actionable. Instead, by adding just a small amount of rigor to our process, we can ultimately be grains of sand on scale slowly shifting business culture towards more thoughtful experimentation and intelligent risk-taking.
Innovation has been a longstanding buzzword in Silicon Valley, and with the rise of concepts like design thinking and circular design from IDEO, and Sprint from Jake Knapp and Google Ventures, many companies have begun the process of integrating innovation as a continual process (as opposed to goal) into their central work processes.
The next natural outgrowth after a few years of hacking away at innovation has become experimentation programs. What may look like the latest business management fad to some, to me looks a lot like the next step in business evolution. For any business dealing with significant scale, it becomes quickly apparent that the experimentation aspect of innovation requires a program all it's own. If you're just now hearing about this, a quick search for "experimentation" in " Jobs" on LinkedIn yielded a wealth of results, many from top companies. It is the next wave of this change.
Great! ... Kinda. The problem with this, like most business fads, is that while it's great so many companies are diving head first into innovation and experimentation, what's being lost in the sauce is often the rigor. I often advise well-meaning business leaders and teams that are devising experiments to run in their business (great 👍) but they aren't taking the relatively small amount of time it takes to put structure to these experiments. As I've discussed previously, innovation and experimentation can be very costly, so it' s important that we are taking intelligent risks, or it will quickly become a financial liability. Many companies (not naming any names) have started innovation programs only to scrap them a few years later. This has mostly to do with the how expensive it can be to run one—unless you're doing it right.
So do yourself a favor (and everyone else in the industry) and devise a structured procedure to test a well-thought-out hypothesis. Make sure you are delineating a clear and repeatable step by step process. Be explicit about your plan to elicit collect and analyze data and feedback, and whether or not you will be running multiple trials or running the experiment multiple times to understand your results. And for all our sake's please have a plan to communicate your lessons learned before you start. Here's a quick checklist I run for through for myself when I create a procedure.
- Is it repeatable? Could someone else understand these simple steps to repeat the process without having to give them a ton of background?
- Is the process testing the right things? Does it test a hypothesis by the same variables I gathered baseline measurements for?
- Is it rooted in a genuine problem statement, or observation about our business performance?
- Is it unbiased? (No, it's not.) But have I gone to reasonable lengths to eliminate bias from the process? (See here for a decent place to start.)
- Do I have a clear and explicit plan for how to review, organize and communicate my findings?
Once you develop a process, be thorough, systematic, and consistent about following it. Be down right religious. Use it to carefully collect new data and judge whether your procedure is effective at testing for the variables you are tracking.
It's not good enough to just try something new and call it experimentation. When we do things like this, we mistakenly give people the idea that experimentation itself isn't worth the investment, when in reality it's only not worth it if we aren't making our learnings actionable. Instead, by adding just a small amount of rigor to our process, we can ultimately be grains of sand on scale slowly shifting business culture towards more thoughtful experimentation and intelligent risk-taking.
