GapRadar: the questions your site never answers
GapRadar reads your own first-party signal — every question visitors put to your on-site chat and voice. It is two things at once: the questions I track, and an honest map of what your content fails to answer. Where Citation shows demand reflected off other sites, GapRadar is the live signal coming straight from the people already on yours.
Each visitor question travels the same short path before it becomes a theme you can act on. Nothing is invented along the way — I only label, group, and count.
Every question is stored with the model's own self-grade — answered or not. The gaps pass then labels each question (the same label is reused for same-topic questions), groups the labels into themes, and rolls the answers up into a coverage score. Themes are replaced wholesale each pass — current themes only, no history — roughly weekly.
A theme's coverage is simply the share of its questions your site's content actually answered. Your headline coverage score is that same share across every theme — answered asks over total asks, 0–100. It carries the usual health bands: go at 80 and up, wait from 50 up to 80, alarm below 50.
| Theme (illustrative) | Asked | Answered | Coverage |
|---|---|---|---|
| Pricing & Plans | 24 | 22 | |
| Integrations | 18 | 11 | |
| Data Residency | 9 | 2 |
Themes and figures above are made up to show the shape — your real themes come only from questions your own visitors ask.
Read top to bottom, the low rows are the work: Data Residency is asked often enough to matter and answered almost never. That is a content gap with a customer already standing in front of it.
The demand log is raw visitor input, so I keep only what the analysis needs and only for as long as it is useful.
Obvious direct identifiers — email addresses and long digit runs such as phone or card numbers — are masked before a question is stored, never after (GDPR data minimization). Recorded questions are then pruned after a retention window, 90 days by default.
GapRadar is the bridge from what people ask you to what AI should cite you for. It does not sit on its own:
If you want the upstream view, the chat & embed page explains where these questions originate, and Readiness covers how a page earns the right to be the answer once you write it.