Memory
Feedback Synthesiser
A pile of reviews, tickets and survey replies turned into two honest rankings: what people say most, what hurts most, and your customers' own words as proof.
Reads your operating system before it starts. Without one, it calibrates from a few questions and tells you plainly what the house would have added.
The task
Three months of reviews, support tickets and survey replies have piled up, and you need to know what to fix first before the next product decision.
By hand
You open the ticket queue and start a tally in a spreadsheet. Two hours in, the categories have drifted: 'shipping' and 'delivery' are separate columns, half the survey replies fit nowhere, and the angriest comment from last Tuesday has quietly become the top theme because it's the one you remember.
By the end you have a list ranked by nothing in particular, one loud complaint mistaken for a common one, and no quote you could actually put in front of the team without going back to find it.
With the specialist
Three answers
The raw feedback, the product area, and the timeframe. If the paste is already scoped and dated, nothing is asked at all. Where your decision ledger is installed, any theme that pushes against a settled call is flagged as tension, not passed off as a neutral finding.
- QThe raw feedback, pasted or as a file
- QThe product area it concerns
- QThe timeframe it covers
The ship gate
Every count is reconciled to the specific items it sums, every quote is checked word for word against the source, and the two rankings are confirmed to be genuinely separate judgements, not one list printed twice.
- Two genuinely separate rankings, by frequency and by severity; identical lists mean severity was never actually judged
- Every count reconciles to specific items in your input; no rounded, estimated or illustrative numbers
- Every quote is word for word from the source, fifteen words or fewer, trimmed only from the ends, never reworded
- Quotes name a source type, never a person; any personal name in the raw feedback is stripped
The deliverable
Two tables: themes ranked by how often they come up, and separately by how much they hurt, each with a short verbatim quote attributed to a source type. One-offs are listed apart, because a single comment is not yet a theme.
Sample output
What the deliverable looks like, on a neutral example.
A quarter of feedback for a small online shop
Ranked by frequency
Ranked by severity
Every count reconciles to the items behind it, and the two lists usually differ in order. One-offs sit in their own list, along with any theme in tension with a call you have already settled.
Most common and most painful are two different lists. Decide which one this quarter is about.
By hand
typically 3 to 4 hours of reading and tallying, with the categories drifting as you go
With the specialist
the feedback pasted in, minutes later two rankings, then your review
Illustrative comparison from the authors' own use; estimates, not measurements.
Feedback Synthesiser: £19, yours to keep.
Buy once, updated free forever. No subscription. Leave your email and you’ll hear the moment it launches.
Runs on your own Claude subscription. Prices are indicative while we finish arguing about them.