Feedback Loop
A feedback loop is the repeating cycle where user input shapes a product change, and that change is observed and fed back in again.
The loop, closed
What is Feedback Loop?
A feedback loop is the repeating cycle where user input shapes a product change, and that change gets observed and fed back into the next round of decisions.
The term originates in systems theory. Norbert Wiener's 1948 book Cybernetics: Or Control and Communication in the Animal and the Machine laid out feedback as the mechanism by which a system regulates itself, using a thermostat and a steam engine governor as the classic examples of a system that reads its own output and corrects course. In product terms, the equivalent reference is Eric Ries's Build-Measure-Learn loop from The Lean Startup: you build something, measure what actually happens, learn from it, and let that learning shape the next build. Ries put the stakes plainly: "All successful startup processes should be geared to accelerate that feedback loop."
Origin is genuinely split into two camps. Engineers trace the term to Wiener and control theory. Startup people trace it to Ries and Build-Measure-Learn. Produck uses the Ries framing, because a feedback loop only earns its name in a product context if it closes, meaning the input eventually changes something a user can feel.
Why it matters for product-market fit
Most teams collect user feedback constantly and still don't have a working feedback loop, because feedback that sits in a spreadsheet or a Slack thread never closes the circuit. It reaches a team, but nothing downstream tells the user what happened to it, and nothing upstream tells the team whether the same complaint is worsening or fading.
This is exactly the gap Produck's loop is built to close. Listen captures the raw signal. Diagnose groups it against what's actually driving churn or delight, not just volume. Decide turns the diagnosis into a prioritized call. Ship closes the loop by shipping the change and, ideally, telling the people who asked for it. Skip any one of those four stages and you have a pipe, not a loop. As we argue elsewhere, most of the standard product-market-fit stack quietly stops at Listen and never gets around to Ship, which is why teams sit on mountains of feedback and still can't say what to build next.
A closed feedback loop is also how iteration becomes deliberate instead of accidental. Iteration without a loop is just changing things and hoping. Iteration with a loop means every change is a response to something real, and every response gets checked against what happens next. And a healthy loop shows up directly in retention: users who see their feedback acted on come back, and users who feel ignored churn quietly, often without ever filing a complaint you can point to.
When it works, and when it doesn't
It works when:
- The loop closes on a visible timescale, so a user who reports something sees a change within weeks, not quarters
- Someone owns turning raw signal into a decision, instead of feedback landing in a channel nobody is accountable for
It falls short when:
- The team collects feedback faster than it can act on it, so the backlog grows and nothing closes
- The loop only runs one direction, capturing input but never reporting back what changed, so users stop bothering to give feedback at all
- Feedback gets diagnosed by volume instead of by impact, so the loudest complaint wins over the one actually driving churn
- The loop measures the wrong thing, tracking ticket counts instead of whether the underlying behavior changed
How to apply it
- Pick one channel where real usage signal already lands, whether that's a support inbox or an in-app prompt, and commit to reading all of it weekly.
- Tag each piece of feedback against a small set of themes rather than treating every item as its own one-off request.
- Once a week, pull the top theme and make an explicit decide-or-defer call on it, and write down why.
- Ship the smallest version of the change that actually tests the theme, not the full-scope version.
- Close the loop by telling the users who raised it that something shipped, even a one-line note.
- Watch whether the same theme keeps resurfacing after you ship. If it does, your fix didn't address the real cause, and the loop just told you that.
Sources
- The Lean Startup | Methodology, Eric Ries, The Lean Startup (2011)
- Cybernetics: Or Control and Communication in the Animal and the Machine, Norbert Wiener, Wikipedia (1948)
