Why taxonomies decay
A feedback taxonomy is built for the business as it was. Then a new app feature launches, a returns policy changes, a delivery partner is added, and customers start writing about things the categories never anticipated. Unless the taxonomy changes with them, new problems hide in "Other" or in categories too broad to act on.
Warning signs
- "Other" or "Unclassified" growing month by month.
- One category holding several distinct problems.
- Categories with almost no mentions for months.
- Teams describing problems that do not map to any category.
- Two people tagging the same comment differently.
A quarterly routine
- Read a sample of "Other" and unclear feedback; propose new sub-themes where a topic recurs.
- Read the largest categories; split any that hold several problems.
- Retire or merge categories with very low volume.
- Update definitions and examples where tagging disagreed.
- Record every change with its date.
Change the right level
- Specific issues: change freely; they should track reality.
- Sub-themes: change a few times a year.
- Themes: change rarely, ideally once a year at most.
Themes, sub-themes and issues: a three-level taxonomy.
Themes sub themes and issues a three level taxonomyProtecting trends
When a sub-theme is split or renamed, reclassify past feedback where possible so before-and-after comparisons stay valid. If that is not possible, mark the change on every chart that crosses it.
Free review snapshot
Want this done for your own reviews?
No commitment. We use your details only to send your snapshot and follow up. Privacy policy
An example
A fashion retailer's "Other" category doubled over two quarters. Reading it showed most were about a new buy-online-pick-up-in-store service, which had no category. A new sub-theme under Delivery and pickup, with history reclassified, revealed that pickup delays were one of the fastest-growing complaints.
Common mistakes
- Never reviewing the taxonomy.
- Changing themes every quarter.
- Not recording changes.
- Leaving "Other" unread.
Common questions
How big should "Other" be?
Small; if it grows, something new is happening.
Who decides changes?
The taxonomy owner, with theme owners' input.
Can AI suggest new categories?
It can propose groupings from unclassified feedback; a person decides.
What about seasonal topics?
Keep them as sub-themes or issues that are quiet off-season, rather than adding and removing them.
Checklist
- "Other" read quarterly.
- Large categories checked for splits.
- Changes dated and recorded.
- History reclassified after big changes.
In short
Expect your taxonomy to decay. Read "Other" and the biggest categories every quarter, change specific issues freely and top levels rarely, record every change and reclassify history so trends survive.
Who owns the taxonomy
Give the taxonomy a single owner, usually in the CX or insights team. The owner runs the quarterly review, keeps the change log, decides on splits and merges with input from theme owners, and makes sure every report says which taxonomy version it used. Without an owner, categories drift as different people add their own.
Share the change log with everyone who reads the reports. A short note such as "From July, Delivery > Pickup is a new sub-theme; earlier months reclassified" prevents confusion when a new category suddenly appears in a chart.
Plan for launches
When the business launches a new product, service or policy, add a placeholder sub-theme before launch so early feedback lands somewhere visible. It is much easier than discovering three months later that the launch's problems were hidden in "Other".
Doing this with Falcon
Falcon builds the theme list from your own reviews, then labels every review against it with a multilingual model, scoring sentiment for each mention and keeping the exact quote behind each label. Texts too short to judge are marked unclear rather than guessed, and low-confidence labels are flagged for a person to check. Negative share by topic, location and month then shows what is rising where, with small samples flagged.