Shipping and specialties
Feedback channels
Every channel lies in a different way. Combine support, in-app, sales and behavior — never let volume from one source set the roadmap.
A launch is a hypothesis entering production. Everything after that is iteration: noticing what users do and say. Teams that treat feedback as a pile of anecdotes ship randomly. The first skill is knowing how each channel is biased.
Never let one source dominate. Power users, angry users and big accounts are over-represented. Volume of complaints is not frequency in the user base.
Concept
Selection bias in feedback means the people who speak are not a random sample of users. Join qualitative themes to analytics counts before you treat a theme as common.
Walking through the lies
| Channel | Good for | Systematic bias |
|---|---|---|
| In-app (widgets, microsurveys, session replay) | Fresh context, UI friction | Over-indexes on power users who click widgets; misses silent churners |
| Support tickets | Broken workflows, severity, wording users use | Biased to people who contact support; loud ≠ common |
| Sales / CS | Deal blockers, enterprise edge cases, willingness to pay | Biased to revenue and logos; can drown PLG signals |
| NPS / CSAT | Trend over time, detractor follow-up | Score alone is almost useless; open text is the product |
| Community (forum, Discord, X) | Passionate edge cases, public reputation | Extremely skewed; great for discovery, bad for prioritization math |
| Behavioral data | What people do at scale | Silent on why; needs interviews to interpret |
NPS asks how likely someone is to recommend you on a 0–10 scale; promoters (9–10) minus detractors (0–6). CSAT asks satisfaction with a specific interaction. Both are lagging mood meters. Treat the verbatim and the follow-up call as the asset; treat the single number as a weak trend line.
NPS is especially dangerous as a prioritization input. A three-point swing can be noise, sampling change or a single outage week. Executives love the number because it fits on a slide. Product teams should love the detractor interview that follows. App-store ratings are smoke alarms: then dig into themes with counts.
Sales-sourced feedback needs a translation layer. "Customer will churn without feature X" might mean a real blocker, a negotiation tactic, or a one-off workflow. Ask: which persona, how many similar accounts, what workaround exists today, and does it show up in support or usage data? One seven-figure account can still justify a build — just label it enterprise exception, not "what the market wants."
Community feature boards help close loops at scale, but they amplify selection bias. Votes without account identity are easy to game and weak as evidence.
Common mistakes
- Prioritizing on ticket volume alone. Loud is not common.
- Treating NPS as a KPI to ship against. Use the verbatims.
- Letting one logo rewrite the roadmap. Label enterprise exceptions honestly.
Check yourself
Support has 90 tickets about notifications. Analytics shows mute rate up and retained muters +8 points. What might that mean?
Mute is a coping mechanism for fatigue, not just disengagement. Volume plus behavior beat either alone.
Sales says a seven-figure account will churn without feature X. Is that "the market"?
Not yet. It may be an enterprise exception. Check persona, similar accounts, workarounds and usage.
Key takeaways
- Every channel is biased. Combine support, in-app, sales and behavior.
- NPS and votes are weak prioritization inputs. Verbatims and counts are the asset.
- Label logo-driven work as an exception, not as market demand.
Further learning
Next, turn themes into numbers in quantifying feedback.
Related
Iteration and feedback loops
Turning feedback, support tickets and data into continuous improvement and kill decisions.
User research
Interviews that avoid false positives, usability tests with five users, and synthesizing findings. Start here for the map.
Product analytics
Designing event tracking plans, reading funnels and cohorts, writing the SQL that answers product questions, and avoiding mix-shift illusions. Start here for the map.