The market feedback loop, closed.
Also called: Voice-of-customer loop · Post-launch insight loop · Continuous discovery · Field feedback system
Capturing structured signal from buyers after launch, spotting the patterns, and ticketing each one into the next product decision.
Launch is the start of the data, not the end of the work. Pull signal from reviews, support logs, returns and the community, look for the pattern that repeats, and route it into a V2 backlog. A shipped product that nobody learns from is a one-off, not a line.
What the feedback loop is
A market feedback loop is the post-launch discipline of treating every customer interaction as a signal worth keeping. The product is in real kitchens now, so the field is telling you things the test bench never could. The loop captures those signals in a structured way, reads them for patterns rather than one-offs, and turns the patterns into decisions you can act on. It is the difference between “we shipped it” and “we shipped it and learned from it”.
The inputs are not exotic. They are the channels you already have: the DTC reviews customers leave, the support emails and messages they send, the returns and the reasons attached to them, and the threads in whatever community your buyers live in. Each one is a flow of raw signal. On their own they are noise: a single grumpy review, one odd support ticket, a handful of returns. The loop’s job is to gather them into one place, watch for the same thing said three different ways, and decide whether it is a fluke or a fault. The output is not a feeling. It is a backlog item with a clear next action against it.
Why a structured loop beats a good memory
The failure mode here is informal. The founder reads the reviews, half-remembers the support questions, has a gut sense that “a few people mentioned the lid”, and never quite acts on it. Three months later the same complaint is still arriving and nothing has changed. A structured loop fixes this by writing the signal down the moment it lands, tagging it, and counting it. Counting is the point: it converts “I feel like people keep asking this” into “eleven people asked this in six weeks”, and the second statement is the one that justifies a design change. The loop is not bureaucracy. It is the memory the business needs so the next version is better than guesswork.
The loop, worked
The clearest way to see the loop is to lay out one product’s signals as they actually arrived. Here is the £149 proofing box’s, the one we ran through to launch, so you can see real signals becoming real backlog items rather than a generic flowchart.
Notice the last row is where the loop closes. Every signal above it has been turned into something a person can pick up and do. A review nobody actions is just a compliment or a complaint; a review that becomes a backlog ticket is the next product taking shape.
- Reviews read once, then forgotten by the weekend.
- Support questions answered, but never counted or tagged.
- Returns refunded with no reason ever recorded.
- The same complaint arriving for months, unactioned.
- Every signal captured in one place the day it lands.
- Patterns counted, so a fluke is told apart from a fault.
- Return reasons coded and traced to batch or design.
- Each pattern a V2 ticket with an owner and a next action.
How it fits the bigger picture
The Market Feedback Loop is the activity that keeps Stage 10 Deliver from being a full stop. The launch put the product in the field; this loop reads what the field says and feeds it back upstream. Its tickets do not vanish into a drawer. They become the input to the next pass through the process, where a V2 is scoped, engineered and shipped. That is what makes the ten stages a loop rather than a line: Deliver ends by handing evidence back to the start.
What it can do
It turns scattered customer signal into a prioritised, evidence-backed list of what to change next, and it tells you which problems are real patterns rather than loud one-offs. It also catches things early: a batch fault flagged by returns data can be isolated and fixed before it spreads across the run, which on a first run of 500 to 1,000 units is the difference between a contained issue and a recall.
What it can’t do
It can’t design the next product for you, and it can’t tell you which requests to honour. The loop surfaces demand; judgement decides what is worth building. A vocal thread asking for a feature is a signal, not a mandate, and chasing every request is how a clean £149 product turns into a bloated one nobody understands. Read the signal, then decide.
See the full 10-stage process →
Try it yourself
List the channels where your buyers already talk to you: reviews, support, returns, social, any community. Against each, name how a signal gets captured and where it lands. Then pick one real complaint from the last month and follow it: was it counted, was a pattern checked, did it become a ticket? Anywhere the trail goes cold, your loop is open, and the field is teaching a lesson nobody is writing down.
Not sure your product is ready to gather field signal yet? Start the Free Sprint → and map the path to a launch worth learning from.
Your feedback-loop checklist
Project notes: the eleven people who asked the same thing
▸ From the notebook · optional reading
Reading the proofing box’s first six weeks of field signal with Dan and Anna Hartley in Stockport, and the question that turned out to matter.
3 min read · click to open
Six weeks after the first DTC units shipped, Dan was pleased. The reviews were warm, returns were low, and the Sourdough School audience had been generous. He told me the product was landing well, and it was. The risk, at that point, was complacency: a good launch is exactly when a founder stops reading carefully.
The signal hiding in plain sight
I asked Anna to do one thing: paste every support message from the last six weeks into a single sheet and tag each by topic. It took an afternoon. When we counted, one tag dominated. Eleven separate people had asked, in their own words, whether the box would still hold 26°C in a cold kitchen in winter. “That’s not eleven worried customers,” I said. “That’s eleven people telling you the same thing.” On its own each message had been answered and closed. Counted together, it was the loudest signal they had.
What we actually did
We split the signals by what they demanded. The knob feel and the lid seal went to the V2 design backlog. The cold-kitchen worry was, on the data, a documentation gap rather than a fault: the box held temperature fine, but nothing told buyers so. A clear note in the manual closed it for almost nothing. The returns cluster we traced to a single batch of Manchester PCBs, isolated by serial range, and took to the supplier rather than the drawing board. And the cooler-mode request from the community went into the backlog as a feasibility study, not a promise.
None of this was clever. It was just written down and counted. Three months on, the V2 scope wrote itself from that sheet, and Dan was scoping a second run of around 3,000 with a backlog built from evidence rather than opinion. The afternoon Anna spent tagging messages is the reason the next product is a decision and not a guess. A loop you actually close is worth more than any single bright idea.
— Deliver stage, project notes, 2026
— Back to the start → The 10-stage loop · V2