Experimental Hygiene

Do you know why WD-40 is called WD-40? Beyond being a warning against putting engineers in charge of branding, the product's name is actually an important lesson for innovation... more on that in a moment.

I just got done hosting the first LaunchPad office hours of the new year. We are interviewing applicants to determine their fit for the program. Each spring quarter, Perry Klebahn and I take 10-15 new teams into the program with the intent to launch real businesses into the world. Over the past 11 years, over 50% of our students' ventures are still in business, have pivoted to another entrepreneurial venture, or have had a profitable exit. Pretty cool stuff.

I'm sure a lot more posts will be dedicated to entrepreneurial topics in the months during which the program runs. This is just a quick note to mention something very important about launching new businesses.

We advocate an experiment-driven approach to bringing new ideas into the world. That means that rather than doing a bunch of market research, we believe in launching in order to learn. Trying things out. Making real offers to real customers. etc.

But importantly, not all experiments are created equal. Some experiments yield enormous amounts of high-quality data, while others are largely a waste of time. What distinguishes the two? One of the topics that came up today was the notion of "experimental hygiene."

Many experiments are sloppily executed and poorly evaluated, if ever. Many founders scramble, frantically doing a bunch of disparate activities, and end up going in circles.

Simply put, a good experiment is clear on two things: what am I trying to learn (ie the dependent variable, or outcome I'm measuring), and what I am doing to learn that (ie the independent variable, or the thing I'm changing). Implicit in the experimentation process is the admission that, "I don't know the answer to this question." More on that later, as it's critical to remember, lest one commission a set of experiments and then abandon them without harvesting the learnings they provide.

For example, one student has a successful product that folks are using repeatedly, but he's having trouble reaching new customers. One important thing for him to learn will be "how can I turn repeat customers into recommenders?" Objectively, he does not know the answer to this question. So should he proceed. Conventional wisdom says, do web research and implement your findings, fingers crossed that they work. (Good luck with that.)

What we'd say is, do your web research, ask others with experience, have a brainstorm, and then (IMPORTANTLY) select 3-5 of the highest-potential ideas and try them out in parallel (over roughly the same time period), and then compare the results of the experiments with each other (ideally along the same dependent variable, or outcome metric). Reflecting on experiments is crucial to closing the loop on the hypothesis-test loop.

Blindly implementing something that might work isn't a great strategy; deliberately commissioning parallel experiments to discover what will work is smart. But you must remember that you don't know the answer, and you're undertaking the experimental approach to potentially discover which answer best fits your situation. "This article outlines a theoretical foundation for why parallel prototyping produces better design results..." if you're curious to learn more about the empirical support for such an approach.

And by the way, all this is not to presume that one of the 3-5 selected ideas will necessarily "work." They very well may not. But 1) they'll provide a baseline of comparison for the next set of experiments, 2) they often lead to better ideas you couldn't have thought of without experimenting, and 3) they are an opportunity to have a data-rich follow-up conversation with the "repeaters" who chose not to become "recommenders". Real decisions are always the best basis for digging deeper.

Which brings me back to WD-40. It stands for "water displacement 40," which is to say that it's named after the forty attempts it took to develop a formulation that successfully displaces water. The team didn't get it right the first 39 times they tried to do it. The astounding thing about that is, they were a well-funded, established team with a clearly defined problem. For most founders getting started, their problem is typically MUCH LESS well-defined than the folks at WD-40; so will it takes more attempts, or less?

This is a helpful frame for folks undertaking experimentation: if results don't come back promisingly, no worries. You weren't running a 400-meter dash! You've just completed the first lap of a much longer race. Don't be crushed by your performance in the 400 (ie if your experiments yield favorable results in the first attempt). Remember the folks at WD-40: you've gained important learning about what DOESN'T displace water! As Thomas Edison famously described his own experiment-driven process of discovery: “I have not failed. I've just found 10,000 ways that won't work.”

Click here to subscribe to Paint & Pipette, the weekly digest of these daily posts.

Previous
Previous

Complimentary Collaborators

Next
Next

We Are All In The Ideas Business