Control the Variable. Not the Curiosity.
We used to experiment ALL. THE. TIME. Before we knew what an experiment even was, and WAY before someone put it on a slide with the word INNOVATION…
…and the first rocket image they found on Google - or in my case, emoji(s) 🚀 🚀🚀🚀🚀, we were doing it constantly. Think about how you learned to walk. You didn't just hurl yourself off the coffee table and hope for the best. Although, to be fair, it sure looked like that - adorbs.
You'd watched everyone around you do this walking thing, so you already had a working hypothesis. “If I straighten my legs and distribute my weight like this”. You had confidence, built from instinct and evidence. You had a clear outcome in mind: mobility - or more accurately, general chaos. Then you tested it. Fell down. Adjusted a variable. Tested it again. Fell again. Confidence, knocked? No. Persistence wins. Nobody wrote you up for it. And you ran. You ran free. (Nooooo, Fennnntooon!)
Life was one big experiment.
Repeatedly dropping this blueberry on the floor - why? How many times will Mum pick it up and hand it back? Just wanna know what happens. Hitting Dad in the face with my tiny, ineffectual, jam-covered hands - why? What's acceptable here? Sequencing my words like this - why? Do they understand me better this way?
And then we got to learn about it as a concept. School handed you a bunsen burner and told you to run experiments. Remember that feeling? The excitement! The sheer empowerment of it! A school handing a child FIRE. (Seriously, there were kids in my school I reeeally would not have). But also a beaker, your hypothesis, and the instruction: go on then, see what happens. It was really just formalising what you'd been doing since you were born - except this time you wrote down what you expected, ran the test, and recorded what actually happened. That's it.
I think it's one of the last times most of us are really tasked with uncertainty as the point, not a flaw to be managed out. You weren't marked on whether your hypothesis worked. You were marked on whether you'd written down what you expected to happen, what happened instead, and maybe even what might work next time.
But most of what we got scored on outside that wasn’t “what did you learn?”; it was “did you get it right?”. Over time, the confidence that counteracted uncertainty curdled into something rather limiting. Soon you start only running the experiments where you’re sure you already know how they end - perhaps because you’ve seen it happen elsewhere. Which is rubbish, frankly. It kills curiosity and the promise of evolution. (I was going to write "innovation" there. See how I didn't?)
Grown now, same fear, where’s my lab coat?
Adulthood, work, all of it - it brings back exactly that same fear of not getting it right.
We call an experiment that didn't go how you'd hoped a "miss." A hypothesis, somewhere along the way, starts getting treated like a commitment on a roadmap - and fair enough, roadmaps need commitments, businesses need outcomes - nobody's arguing with that. The problem is when the test itself gets held to the same standard as the promise. When did we stop treating uncertainty as a normal part of getting to a good outcome, and the disproved hypthesis as something to hide away from view?
When something doesn't go to plan, we hear about post-mortems, root cause analyses, "lessons learned" documents written once, read never, and importantly - those hypotheses are rarely tweaked to be tried again. All useful, in theory. Still you can treat the result as something to explain away rather than something you actually went looking for. We didn't lose the capacity to experiment. Perhaps we lost setting that made it okay to be wrong out loud - and then, mysteriously, wondered where all our boldness went.
Can we sit more comfortably in this discomfort?
Well, yeah. Here’s some tried and tested ways:
Cheap tests before expensive ones. Not every hypothesis needs a roadmap slot and three teams holding hands around it. Some do, maybe. LaunchDarkly's ENTIRE operating rhythm is built on this: they ship 60 to 70 updates a day (!!) treating each one as a small, reversible test rather than a “Big Launch Moment”, which is exactly what lets them fail smart instead of fail spectacularly. I mean that’s truely an extreme but : the cheaper and faster the test, the closer you get to the right answer.
Kill (or is it "unalive" now? 🤔 ) criteria - decide before the result. Not a new idea, and certainly not mine. Stage-Gate found that most organisations have no formal mechanism for stopping a failing project, which is exactly why money and resource keep flowing into things that are already… well dead unalived. "Should we end this?" is SO HARD once you’re inside, making the decision. Hours of work in, emotionally invested up to your eyeballs. Decide the criteria before you're attached, and an unexpected result is just learning and data to adjust. Not a reflection the work you’re doing to get to the answer.
Signals, not confessionals. If the only visible artefact of an experiment is a success story, you've bascially told everyone what's safe to bring back to the team. Retrospectives structured to reward the learning, not just the win, give a "failed" test somewhere respectable to live - a shared knowledge base, say - instead of getting buried in a Slack thread, privately closed down in a Teams meeting, and never spoken of again until someone else tries the same thing and gets told "we've done that already," with no memory of what was actually tried and we learned. [takes breath] And certainly no progress since in getting to the outcome.
Time and budget that's protected, not just promised. Experiments are the first thing sacrificed the moment a deadline tightens. Does your AOP have an allocated resource, time and budget for experimentation? How will you guard it from getting absorbed whenever the "real" work needs a bit more room?
And look - truly, experimentation isn’t just about product shipment. That internal comms breakdown that remains unsolved. The report you still have to generate from several sources every month. Getting proper access to a shared resource without three approvals and a prayer. All of it is an experiment waiting to happen, if you let it. Keep working to that outcome.
Hand me my b—urner
The kiddo at the lab bench wasn't fearless. She knew the result was learning. 🚀 Innovating🚀 by design. A space (physically and metaphorically), and time (…er…), built deliberately so that trying something and learning from it was just what happened there. Keep walking toward the end outcome.
What do you think? Did we become too fearful to be experimental and curious? Is your bunsen burner in reach?
