The whiteboard nobody wanted to erase
Years ago I sat in a room where a junior engineer had pushed a change taking a service down for eleven minutes. Eleven minutes doesn't sound like much until you're the one explaining it to a customer on the phone. The old version of me wanted to find out who signed off on the deploy and make sure it never happened again the easy way: fear.
Instead I asked him to walk the whole team through what happened, on the whiteboard, in front of everyone. Not as a trial. As a lesson. He drew the sequence of events, the assumption he'd made, the test he'd skipped because "it always passes anyway." Twenty minutes later, three other engineers admitted they'd skipped the same test for the same reason.
This whiteboard stayed up for a week. People walked past it and looked at it the way you'd look at a scar. It worked because it turned one person's mistake into everyone's lesson, instead of one person's secret.
We say we want innovation. We punish the thing producing it
Every leadership team I've worked with says they want more innovation, faster experimentation, people willing to take smart risks. Then someone's experiment fails and the org chart quietly rearranges around whoever gets blamed for it.
You don't get both. Innovation is a string of attempts, and most of those attempts don't work. If your team believes a failed attempt gets logged against them personally, they stop attempting anything risky. They stick to the safe, proven path, and dress it up under the name "discipline."
Research backs this up more bluntly than I expected. A 2025 survey of 2,000 UK workers by Mental Health First Aid England found 45% of employees don't feel safe raising mistakes or flagging risks to their employer. Not "don't want to." Don't feel safe. Nearly half your workforce sits on information you need, and silence still feels safer to them than honesty.
Not a training problem. A leadership problem, and it starts with how the last mistake in the room got handled.
Google already ran this experiment for you
Google spent years studying what separates its highest-performing teams from the rest, in a project internally called Project Aristotle. Of the five dynamics identified as driving team effectiveness, psychological safety is listed first, ahead of dependability, structure and clarity, meaning, and impact.
Psychological safety, in Google's own words, comes down to a team member's sense of the consequences of taking an interpersonal risk. Speaking up when something's wrong. Admitting you broke something. Proposing an idea which might flop in front of the whole room.
None of this happens in a culture where the last person who made a mistake got quietly sidelined. People notice. They watch what happened to the last person who admitted fault, and calibrate their own honesty against it.
I've spent a lot of my career studying bad bosses, and my own research found 99.5% of people surveyed have had at least one. Punishing mistakes instead of learning from them is one of the fastest ways to earn a spot on this list.
What "celebrate failure" looks like in practice
I want to be direct about something, because "celebrate failure" gets treated like a poster in a break room and nothing more. It doesn't mean throwing a party every single time someone screws up. It doesn't mean removing consequences for negligence or repeated carelessness. It means something narrower and harder to pull off.
Separate the mistake from the person. The engineer who skipped the test wasn't lazy. He'd inherited a habit from a codebase where the same test flaked constantly and everyone had learned to ignore it. Fix the system, not only the individual.
Make the retelling safe. If someone has to defend themselves the moment they open their mouth, they'll give you the version protecting them, not the true one. You need the true one, every time.
Ask what the mistake taught the team, out loud, in front of the team. Private feedback fixes one person's understanding. A blameless, public review fixes the system everyone else is still standing in.
Watch what gets rewarded. If your top performer never takes a risk because they only ship the safe stuff, and your most promising experimenter gets quietly passed over because a swing didn't land, you've told your whole team what you value, no matter what the values poster says on the wall.
Ask the second question before the first one. The first question after a mistake is almost always "who did this." Ask "what did we learn" first instead, out loud, before anyone gets near blame. Order changes everything about how safe the room feels.
What the Army taught me about mistakes
I served in the Army before ever leading a team in an office, and the difference in how mistakes got handled stuck with me. In training, an after-action review pulls apart everything going wrong, by name, in front of the whole unit, within hours of it happening. Nobody treats it as an attack. It's how you get better before a mistake matters more than it did in a drill.
Business culture, by contrast, tends to bury the mistake and hope nobody brings it up again at the next performance review. We call this politeness. It's closer to cowardice, and it costs teams far more than a bruised ego ever would.
An after-action review works because everyone in the room already agrees on the goal: get better before it counts for real. Bring this same agreement into a design review, a sprint retro, or a product launch post-mortem, and the tone in the room changes almost immediately. People stop protecting themselves and start protecting the mission instead.
The teams which innovate are the teams which fail out loud
Here's the part taking me the longest to believe: the teams I've watched ship the most interesting work aren't the teams with the fewest mistakes. They're the teams where mistakes surface fast, get discussed honestly, and get treated as data instead of evidence in a trial.
Few mistakes usually means few attempts. Few attempts means little learning. Little learning means you're standing still while a competitor with a messier, more honest process walks past you.
I wrote before about why nobody trained me to lead and how much of leadership involves undoing bad habits inherited from bad bosses. This is one of the biggest ones. For years I assumed a clean track record meant a strong team. Usually it means a scared one, quietly hiding its worst weeks from you.
This whiteboard stayed up because the team wanted it there. Nobody erased it out of shame. Eventually someone wiped it clean because the habit it represented had already changed, not because the memory was uncomfortable.
Ask yourself this
Think about the last mistake someone on your team told you about. What happened next? Did they get thanked for surfacing it fast, or did the room go quiet? The answer tells you more about your culture than any values statement ever will.
If you want a team which innovates, you don't need a bigger budget or a new framework. You need to make it safe for someone to say "I broke it" out loud, and mean it when you say thank you for telling me.