The key to scaling feedback ops is to stop handling feedback one item at a time and handle issues instead. This page covers how issues are clustered and how to use them.

What is an issue

An issue (topic) is a cluster of feedback that's all about the same thing. An issue like "coins not credited after top-up" might aggregate dozens of pieces of feedback from the App Store, Google Play, and email — worded differently, but all pointing at the same problem.

With issues, your workload drops from "2,000 pieces of feedback" to "20 issues." The unit you judge and act on is the issue, not the individual review.

Clustering is automatic

Once feedback arrives and is understood by AI, similar items are automatically merged into issues. You don't need to tag and categorize by hand — new feedback joins an existing issue, or forms a new one when enough of a kind show up. Each issue carries: feedback volume, sentiment breakdown, priority, owning feature module, and trend.

Two views

At the top of the board you can switch between two views:

  • Grouped by issue (default): similar feedback is merged into issues — good for seeing the whole picture and working issue by issue.
  • Detailed feedback list (item by item): no merging, each item shown on its own — good when you need to read every original.

Drilling down from an issue

An issue isn't a black box. Click into an issue (L2) to see the feedback it aggregates and its trend; click a single item (L3) for the feedback detail — the original and its translation, plus the AI's classification reasoning. Clustering results can always be drilled into to verify, and reclassified by hand. For the full explanation of the three-level drill-down, see Reading the Board.

Issues are roadmap input

Sort issues by feedback volume × severity × trend and you get an evidence-backed pool of product candidates. For how to turn issues into roadmap input, see the blog post Turning Voice of the Customer into Roadmap Input.