Voice of customer: keeping the support → product loop alive without Slack screaming
Every SaaS has the same problem: support sees the pain first, product hears about it three months later. Here's a loop protocol that survives 12 months.
The voice-of-customer (VoC) loop dies of two diseases: too many unsorted signals drowning product, or too much support-side filtering that only surfaces support's own priorities. The right protocol sits in the middle.
Three distinct channels, three cadences
- "Urgent product bug" channel:instant escalation to the right squad. Strict criteria: reproducible, > 5 customers impacted, no workaround. No debate, no triage.
- "Weekly pattern" channel: written synthesis, sent every Friday. 3 to 5 patterns max. Format: volume, examples, cause hypothesis.
- "Quarterly deep dive" channel: 90 minutes, support + product + design. Broad topic, quantitative data, roadmap decisions.
The pattern format that reads in 30 seconds
A useful weekly pattern fits in 4 lines:
- Observed symptom: short sentence, concrete examples.
- Volume: conversations / 7 days, trend vs last week.
- Estimated cost: average duration × volume.
- Hypothesis: what support thinks, not what it knows.
Golden rule
Support escalates observations, not solutions. An escalation that already contains the solution short-circuits product thinking and usually triggers reflex rejection.The acknowledgment mechanism
The #1 cause of VoC loop collapse: support escalates, product doesn't respond, support stops escalating. Product-side minimum viable: a status per pattern (accepted, rejected with reason, under investigation) within 14 days. Five lines is enough.
A review cadence that forces decisions
- Friday: support synthesis → shared list.
- Following Monday: 30-min product triage, status set.
- Every 90 days: retrospective on accepted patterns. Which ones actually shipped?
Without the retrospective, the system devolves into an accounting of escalations with no impact. That's the step that saves the loop at the 6-month mark.