Feedback Loops

Deliberate practice named immediate feedback as one of its non-negotiables. This chapter is about that mechanism on its own — because feedback is the steering wheel of improvement. Reps without feedback are just movement; reps with it are direction. And the single biggest lever you have isn’t getting more feedback, it’s making the loop faster.

The principle: the speed of the loop sets the speed of learning

Improvement is a control loop: act, observe the result, adjust, act again. The tighter that loop — the less time between an action and the information about whether it worked — the faster you learn. This is one of the most general principles in nature and engineering, and it shows up under many names:

• PDCA (Plan–Do–Check–Act), W. Edwards Deming’s cycle for continuous improvement.
• Kaizen, the Toyota Production System philosophy of many small ongoing improvements rather than rare big ones (Masaaki Imai popularized the term).
• The OODA loop (Observe–Orient–Decide–Act), John Boyd’s model from aerial combat — where the pilot with the faster loop wins by acting inside the other’s reaction time.

The common thread: a fast, small loop beats a slow, big one. A daily two-minute check beats a quarterly review, because by the time the quarterly review arrives you’ve already grooved three months of uncorrected error. Tight loops catch mistakes while they’re small and cheap to fix.

There’s a second reason fast loops win, beyond catching errors early: attribution — figuring out which action caused which result. When feedback arrives right after you act, the link is obvious. Wait a month and a dozen other things have changed, so you can no longer tell what worked from what didn’t — the lesson is in the data, but you can’t read it. A tight loop doesn’t just correct faster; it corrects more accurately, because cause and effect are still close enough to connect.

   SLOW LOOP                       FAST LOOP
   ─────────                       ─────────
   act for months                  act
   then review                     observe immediately
   error compounded the whole time adjust
   few, large corrections          act again
                                   many, tiny corrections

To run a loop you need something to observe, which means you have to measure. You don’t manage what you don’t measure — not because numbers are sacred, but because memory lies and a metric doesn’t flatter you. The metric doesn’t have to be fancy: a tally, a time, a yes/no, a recording. It just has to be honest and to reach you fast.

A simpler way to see it — you can’t fix throws you never see land

Imagine throwing darts at a board in a pitch-dark room. You could throw all night and never improve, because you never see where a single dart lands — there’s nothing to adjust toward. Feedback is turning the lights on. And the speed of your loop is how often they flash: a slow loop is one flash a month; a fast loop is the lights left on, so you correct after every throw. Same effort, but only one of them is actually aiming.

The common trap — feedback that makes you worse

Feedback isn’t automatically good. A classic meta-analysis (Kluger and DeNisi, 1996) reviewed hundreds of studies and found that more than a third of feedback interventions actually reduced performance. The pattern: feedback aimed at the task (“your timing drifted in bar 12, here’s the fix”) helps; feedback aimed at the self (“you’re bad at this,” or even effusive “you’re so talented!”) pulls attention to your ego and away from the work, and can hurt. So when you seek or give feedback, keep it specific, task-focused, and actionable. “How do I improve this?” beats “Was that good?” — and “the intro dragged” beats “you’re a weak speaker.”

The practice: instrument one skill

Pick one thing you’re trying to improve and build a real loop around it:

1. Choose one honest metric. What’s the simplest signal that tells you whether you’re improving? Words per minute, errors per session, a 1–10 self-rating right after, the number of replies your message got. Pick one; don’t drown in dashboards.
2. Shorten the loop. Find the slowest part of your current feedback and speed it up. Don’t wait for the monthly meeting — record today’s attempt and watch it tonight. Don’t wait for sales figures — ask one customer now.
3. Ask better questions. When you seek feedback from people, make it specific and task-focused: “What’s the one thing I should change about the opening?” gets you a usable answer; “What did you think?” gets you “it was good.” Make it easy for them to be honest.
4. Run a short, regular review. Once a week, spend ten minutes looking at your metric and your notes: what worked, what didn’t, what one thing to change next week. This is kaizen at personal scale — small, steady, never skipped.
5. Close the loop. Feedback you don’t act on is just noise. Each cycle should end in one concrete change to the next rep.

How to tell it’s working: the gap between making a mistake and noticing it keeps shrinking. At the start you find out you were off course weeks later, usually from someone else. As the loop tightens, you catch it yourself — the same day, then the same hour, then mid-rep. The day you can feel an error as it happens and adjust on the very next attempt, the loop has become part of how you work, and improvement stops feeling mysterious and starts feeling steady.

In real life — the cook who tastes as she goes

Two people make the same soup. One follows the recipe exactly, never tastes it, and serves it — discovering at the table that it’s under-salted, with nothing left to do. The other tastes it every few minutes: a little flat, add salt; too sharp, a pinch of sugar; taste again, adjust again. By the time it reaches the table it’s dialed in. Neither cook is more talented. One ran a tight feedback loop — taste, adjust, taste again — and the other flew blind until it was too late to steer.

Try this today — ask one specific question

Find one person who’s seen your work and ask a single, narrow, task-focused question: “What’s the one thing that would most improve this?” Resist defending or explaining. Just collect it, thank them, and change one thing. You’ll get more useful signal from that one question than from a year of vague “looks good”s.

The challenge

This week, instrument one skill with a tight loop: pick a metric, observe it daily, and run a single ten-minute weekly review that ends in exactly one change. Compare how fast you improve to your usual rate. The point you’ll feel: improvement isn’t mysterious. It’s a loop, and you can make it spin faster.

How it connects

Feedback loops are the other half of deliberate practice — practice supplies the reps at the edge, feedback supplies the correction. The loop also runs on attention (you can only act on what you actually notice) and on the calm mind (defensiveness is the enemy of hearing feedback clearly). And the weekly review is a small, structured version of reflection — turning raw experience into a deliberate adjustment.

Reflect

1. For a skill you care about, how long is your current feedback loop — minutes, weeks, or never? What’s the slowest part you could speed up?
2. When was the last time you got feedback that genuinely changed what you did next? What made it usable?
3. Be honest: do you seek feedback, or avoid it because hearing it stings? What does that avoidance cost you?
4. Think of feedback that once made you worse, not better. Was it aimed at the task or at your ego?
5. What’s one honest metric you could start tracking this week for something you want to improve?

Show reflections

1. Many people discover their loop is “never” or “far too slow.” Identifying the slowest segment — waiting for a review, for results, for someone else — usually reveals the single highest-leverage fix.
2. Usable feedback tends to be specific, timely, and about the work. If you can’t recall any, it may be because you’ve mostly received vague praise, which feels nice and teaches nothing.
3. Avoidance is common and usually ego-protective. Naming the cost — months of uncorrected error — reframes feedback from a threat to the fastest available shortcut.
4. Feedback that targets the self (“you’re not good at this”) wounds without informing. Recognizing the task-vs-ego distinction lets you extract the useful part and discard the sting.
5. A good metric is simple, honest, and fast to read. If you’re tempted to track ten things, that’s a sign you haven’t yet found the one signal that matters most.