Voice-of-Customer (VoC) Synthesizer

Example output

Voice-of-customer synthesis — Bonvale cold-brew concentrate, 60-day support and review data (85 items)

Escalate first — before anything else in this report
Two mentions of "got sick after drinking this, not sure if related" require immediate escalation to your quality and food-safety team today, regardless of how small the count is relative to your other themes. Two reports of a claimed illness is not a large sample, and the respondents themselves are uncertain about causality, but any illness claim tied to a consumed product carries a severity weight that overrides frequency ranking. Do not wait for this theme to grow before acting; pull the batch/lot information from both tickets immediately if you haven't already. This is a "verify with your food-safety and quality team now" flag, not a synthesis finding to sit in a report.

Ranked themes (frequency and severity, excluding the escalation above)

1. Cap seal / leak issue (high frequency, high severity)
"Bottle leaked in my bag" and "cap doesn't seal right" appear to be describing the same underlying packaging issue from two different symptom angles. This reads as a genuine pattern, not isolated incidents, given how specific and mechanically consistent the complaints are (leak plus seal, not leak plus taste, which would suggest unrelated causes). Recommend routing to packaging/quality as a probable cap or seal specification issue, and pulling the actual ticket count for this specific theme versus your other 83 items to size it precisely; the two example phrases you gave me are illustrative, not the full count.

2. "Tastes burnt this batch" (frequency to be confirmed, moderate-to-high severity)
The phrase "this batch" is a specific and useful detail; it suggests respondents themselves are attributing the issue to a batch variation, not a permanent recipe problem. This is worth cross-referencing against your production batch/lot records for the 60-day window: if these complaints cluster around a specific date range or lot number, it points to a process or roasting variation rather than a formulation problem, which changes who owns the fix (production/QA versus R&D).

3. Portion/size request: "wish it came in a bigger size" (frequency to be confirmed, low severity, high opportunity)
This isn't a complaint; it's an unprompted demand signal for a larger format. Distinct from the two issues above, this belongs with product/innovation, not quality, as a possible line-extension input rather than something to "fix."

What's praise, not complaint
"Love the strength" appears alongside the size request, suggesting your core product concentration and flavor intensity are landing well; worth explicitly protecting as you investigate the burnt-batch theme, so any process fix doesn't accidentally dilute the strength that's driving the positive mentions.

Isolated versus pattern read
With only the phrases you've pasted as examples rather than the full 85-item breakdown, I can characterize the leak/seal and burnt-batch themes as directionally patterned based on how specific and mechanically consistent the language is, but I cannot give you a precise percentage of the 85 items each theme represents without the full data. Paste the complete set of 85 items (or a tally by theme) for an exact frequency count before this goes into a quality review deck.

Suggested owners
- Cap/seal issue: packaging engineering plus quality.
- Burnt batch flavor: production/QA, cross-referenced against batch records.
- Illness mentions: food safety, immediately, independent of this report's timeline.
- Bigger size request: product/innovation, as a future-consideration input, not urgent.

One-page VoC summary
Of the themes in Bonvale's last 60 days of support and review data, two require action ahead of any prioritization exercise: the illness mentions (escalate to food safety today) and the cap/seal leak pattern (route to packaging quality this week). The burnt-batch flavor theme likely traces to a specific production batch and should be cross-referenced against lot records. Separately, unprompted requests for a larger size are a positive demand signal worth flagging to product, not a problem to solve.

Want me to draft the food-safety escalation note for the illness mentions, or build the full frequency table once you paste the complete 85-item dataset?