Without Feedback, We're Just Guessing

It's been a bit since I last posted. Focused on building, versus writing and battled back and forth internally with my own thoughts about AI generated vs AI edited vs AI assisted vs Human only content. Truth is, I believe the majority of us working with AI have at least some assistance from AI tools in a generative capacity for our workflows and content today. Going forward, for my articles and content, I'll include a sign off (it's one of my hard rules for using generative AI) that indicates the level of AI involvement.

After a summer pause, it's time to pick my writing back up and the next in the series is about feedback loops, which have become even more visibly important in daily workflows (with and without AI).

So... Back to it.

You've built clarity and defined ownership. Your team has established decision rights, guardrails, and is building trust yet sometimes may still feel like you're flying blind (or at least not fully "on it" all the time).

Yet, without a feedback loop none of those systems actually tell you whether they're working. And I don't mean the annual review kind, or the "my door is always open" kind. I mean the operational kind, the kind that you truly feel like you know if something is working, not working, on track, off track, and can get a hypothesis as to why.

How Fast You Find Out Shapes What It Costs

How fast you find out about a problem can have more to do with what it ends up costing you than how big it is. Two stories from the same stretch of time in my career show what I'm referring to.

The first one was fast. We had a fire at one of our facilities, it started with a hot ember smoldering in a trash can overnight. I found out from a phone call early am while sleeping and the first thing I did was ask if everyone was safe, thank goodness they were. Within a few minutes, at 5-something in the morning, we had a clear communication process and feedback channel going for status updates. From there, vendor quotes came in as they landed, the insurance claim was in motion the same day, and we started deciding where else we could line up to work out of. We didn't have a full facility fire playbook at the ready (I mean, who does? - especially in an early stage growth company), we focused on communication and made sure everyone who needed to know, knew, without every update running through me or the senior leader on site. We were back in production in a couple days and full production within a week or so. Not ideal... but no doubt a lot better than it could have gone.

The second one I only heard about after it was over, and I touched on it in the trust article. A team member walked out of a production meeting. I wasn't in the room. I found out on Slack and offered to help. He left with a clear end date and a paid transition.

The fire got fixed fast, because everyone could see what was happening as it happened. The walkout reached me secondhand, once it was already over.

Crisis Can Build Infrastructure (If You Let It)

A crisis strips all the noise and overthinking options away, and has a way of forcing you to focus immediately on what's important to solve the root issue. Crises demand building feedback loops that provide clear and concise communication about the things that matter and nothing else. Ironically, these are the ones that probably should have been built all along.

When we were close to going under (the story from my trust article), the daily standup became non-negotiable: every morning, on what was actually true that day and what had changed overnight. When you're that close to the edge, yesterday's numbers are already out of date, but the lessons from yesterday's work can still be applied today.

What I'd point to is what happened after. Those standups kept going when things stabilized and the rhythm we built to survive turned into the rhythm that kept the next rough patch from getting anywhere near as bad. We only had to decide to keep it. As "boring" as some of them seemed, when we leaned into them they always provided positive results.

Donella Meadows wrote about this in Thinking in Systems (worth the read if you haven't). She describes two kinds of loops: balancing loops that pull a system back toward where it should be, and reinforcing loops that snowball, for better or worse. Many companies have plenty of the snowball kind and often let the balancing feedback loops drift as they are easier to look at as "not that important" when things are going "ok" or even when they are going well. Nobody plans it that way. It just happens when you're building fast or in a position of comfort.

The cut power cord story from the guardrails article is one of the pull-back kind. An operator on our team found cut cords across multiple tools, an obvious electrical and fire hazard. Our leadership team set up a dedicated safety channel that same day for folks to post in (pics, text, voice, whatever worked and was easy). Nothing fancy. Just a place where anyone could post what they found and everyone would see it right away... which was the whole point and led to a safer workplace.

Make Honest Feedback Easy, Then Close the Loop

Most leaders I know say they want feedback, including myself. Not all of us build the mechanisms that produce it, and early on in my career I'd have been in that group too.

I said in the trust article that "my door is always open" isn't a real pathway, and I still believe it. It asks the person walking through the door to bring the courage, the political capital and the belief that being honest won't cost them anything. That's a lot to ask of anyone, especially someone new. I've had the open door. I've also watched good people walk right past it so to speak.

The 1 to 10 question worked because it asked for none of that. A new operations lead, first week, and we asked him to rate the operation. He said 6. A specific question gets a specific answer. "How's everything going?" mostly gets "good," or something like "it's another day." Neither of which is helpful to create a healthy culture focused on goals together.

From then on we asked it consistently, at random intervals to make sure we were capturing good data, and posted the answers where the whole team could see them. That 6 became a starting point we could watch move to a higher result (above an 8).

One rule I've written about before belongs here too: don't filter what reaches me unless it's outright profane (and even then, give it to me straight TBH). I want the raw signal. Once people start sanding the edges off before it gets to leadership, the whole thing stops working, and nobody tells you it stopped (this can be combatted by getting to the gemba frequently to observe and understand the work with those actually doing it, more on that in a future article).

Compare that with what a lot of organizations run, an annual engagement survey that comes back months after anyone filled it out and has been reviewed with some statistical modeling. By then the people who were frustrated have left or stopped caring. Perhaps I'm a bit too crass with this - running an annual process is a good start, yet the key here is that the feedback loop cycle should match the goal of your transformation timeline and horizon. Larger organizations take longer for change to be adopted, no doubt. That said, I believe incremental change daily will outpace annual strategic cycles and result in better culture and performance.

Whatever rhythm you pick, the loop has to close. If someone raises an issue and nothing changes, they stop raising issues. When a change can't be made, it's our responsibility as leaders to close the loop with the team and let them know what the decision is and in most cases why it was made. This helps build trust, even when the team or team member may not agree with the decision. A loop that never closes breeds a culture of people filling in their own blanks and leads to thinking that speaking up is pointless, and that lesson sticks around a lot longer than whatever you were trying to fix.

Feedback and Trust Build Each Other

Feedback needs trust to work. And trust needs feedback to be built. They go hand in hand. A 6 out of 10 only shows up when people believe a 6 will be treated as useful and not held against them. When I get a 6 from a team member and use it as a starting point, and the team sees something change because of it, the next number tends to be a little more honest (up or down - either way, more truthful). Same deposits I wrote about in my trust article.

I get that these can seem like simple, no-brainer basics. That's exactly why they get overlooked. They're repetitive, a bit boring, and easy to label as unnecessary for high performance, which is simply not true. Surgery is about as high-performance as work gets, and in 2009 a World Health Organization study led by surgeon Atul Gawande had operating teams in eight hospitals, from Seattle to New Delhi, start using a basic 19-item checklist around every operation. Deaths dropped from 1.5% to 0.8%, and complications from 11% to 7%.

Fighter pilots are another good example. One of my life-long friends is a pilot, and they have some of the most intense training there is, yet they still rely on their team's feedback loops to stay sharp. An F-15C pilot and an Air Force lawyer, writing about what military lawyers can learn from fighter debriefs, describe a 50-minute flight with 6 to 10 minutes of actual dogfighting, followed by a debrief that can run upwards of four hours. The flight lead might be a lieutenant debriefing a colonel, and the lead debriefs their own decisions too. As one Air Force colonel put it, "there is no rank in the debrief." The research backs this up beyond the cockpit: a 2013 analysis of 31 studies found that properly run debriefs improve individual and team performance by roughly 20 to 25%.

Every other piece of this series needs these loops too. Without feedback, clarity slowly turns into assumptions, ownership can drift into isolation, decision rights morph into little fiefdoms of power or fear, and guardrails can turn into the bureaucracy they were meant to prevent. Trust is the one in a corporate environment that fades quietly (and doesn't want to be addressed), until you realize you've stopped actually hearing from your team.

How to Start Building Your Feedback Infrastructure

If you only do one thing, start here and start now... the next time something goes wrong and you spin up an emergency check-in, don't tear it down when the crisis is over. Ask whether it should stay, how it should evolve, and what you (and your team) learned from it - then do it. Usually you'll find that some form of the crisis communication channels should stay, and you've already built them, so use them well.

Then a few more you can work on:

• Measure the gap between a problem happening and the right person knowing about it. A production issue, a customer complaint, a team conflict... how long does each one take to reach someone who can act on it? How can you shorten that timeframe? AI is already shortening some of these gaps for me. I'll get into how at the end of this series.

• Ask one specific question on a regular schedule, somewhere people can see the answers. Something like "Are we on track or off track for the week ahead, and why?" asked every Friday (or Monday) will tell you more than any annual review.

• Close the loop out loud. "Last week you rated us a 6. Here's what we changed. Where are we now?" People keep talking to you when they can see something happened.

A Short Self-Test

A few questions worth sitting with (and I'd love to hear the ones you use in addition to these or adjustments to make to them):

• How long does it take a problem on the floor (in prod, etc.) to reach the decision maker? Is that fast enough?

• What recurring question(s) do you have in place that gives you a number, not just a feeling?

• The last time someone gave you a 6 out of 10... what did you actually do with it?

• When people raise issues (or praise), do they see something change?

• If your best team member quit tomorrow, would you already know why? If so, why aren't you addressing it ahead of time?

Pick the biggest gap and build one small thing for it this week.

Up next: I'll share my thoughts on what I'm calling Zone 2 Leadership... the steady, sustainable pace the best leaders I know actually work at.

This is Article 6 of a 10-part series on how to get quality work done at scale. Each article builds on the last. If you're just joining, start with Article 1: Clarity Is a Leadership Obligation.

AI involvement: I wrote the opening by hand. The rest was drafted with AI, then reviewed and edited materially by me through 14 versions before publishing.

Cheers,
Elton