Skip to main content
Zudo reads the email threads it syncs and groups the questions customers actually ask. The result is a ranked list of what your inbox keeps having to explain — which is usually a documentation gap, an onboarding gap, or a product gap wearing a support costume.

​
How it works

The analysis runs across the threads Zudo has synced from Gmail and Help Scout. It clusters questions by meaning rather than wording, so “how do I export this” and “is there a CSV download” land in the same group. Each recurring question carries:
  • How often it comes up — the number of threads in the cluster
  • Which accounts asked — so you can tell a broad confusion from one account’s confusion
  • Example messages — the actual wording customers used, which is usually better than yours

​
Running the analysis

Start it from the emails section. It processes your thread history in the background and can take several minutes on a large mailbox — you will see progress while it runs, and you do not need to keep the page open. Re-run it periodically rather than continuously. The useful signal is the shape of the last few months, and that does not change hour to hour.

​
What to do with the output

The point is not the list. A question that shows up thirty times is telling you one of three things:
  • The answer exists but is unfindable — fix the docs’ navigation or naming, not the answer
  • The answer does not exist — write it, then link it from wherever the question gets asked
  • The answer is “you can’t” — that is a feature request with thirty pieces of evidence attached, and it belongs in a request
That last case is the one worth watching for. A recurring question that is really a missing feature has already done the work of proving demand.
Recurring questions are computed from email content and are not currently exposed over the API. The underlying threads and messages are — see conversations.