Skip to main content
/dashboard/sources answers a different question from the main dashboard. Not how visible am I, but where are these answers getting their information — and which of those places cover a competitor and not you. It uses the same window, the same confidence intervals and the same suppression rule as everything else.

​
Cited domains

For each domain the engines cited in your window, the table shows how often it was cited, as a rate with its own n and interval.
Citation rate = answers in which this domain was cited ÷ answers whose sources were captured.
That denominator is the careful part. It counts only the answers where rekkal actually recorded a source list. An answer whose sources were never captured contributes to no domain’s numerator, so counting it in the denominator would report every domain the answer really did cite as “not retrieved” — quietly depressing every rate on the page. This is also why the sources view can legitimately say “answers were collected, but their sources were never captured” while the main dashboard shows perfectly real numbers for the same window. Both are true.

​
How domains are classified

Three labels, each read off something rekkal already knows rather than guessed at:
rekkal refuses two guesses here, deliberately.A competitor domain that has not been corroborated is never used for attribution — a model’s guess at which company owns a domain is how a French school gets attributed to a German company. And there is no publisher taxonomy: rekkal will not tell you a domain is “a news site” or “a review site”, because nothing in the data supports that and inventing a category is exactly the confident guess this product exists to refuse.

​
Gaps

A domain is a gap when, among the answers in this window that cited it, at least one named a tracked competitor and none named you. That is a concrete, checkable statement: this source is feeding answers about your category, those answers name a rival, and they do not name you. It is the closest thing rekkal produces to a to-do list. Two rules about which domains can appear:
  • Your own domain is excluded. Your site is not somewhere to go and get covered.
  • A competitor’s own domain is kept, and labelled. Suppressing it would be a judgement call better made by you than by rekkal.
An empty gap list is reported precisely too: “no competitors tracked”, “nothing measured”, and “every source that named a rival also named you” are three different claims, and only the last is good news.

​
What is on screen

Both tables show the top 25 rows and count the rest underneath. It is a display cut only: every rate, sample size, interval and change on the rows shown was computed over the whole window before the list was cut.

​
Not available yet

You can read the answers behind your visibility figure — that card links to them, with every mention of your brand highlighted. There is no equivalent for a citation yet: nothing on this page opens onto the answers a given domain was cited in.
Turning a gap into a specific recommended action is not built yet. The gap table tells you where you are missing; it does not yet tell you what to do about it.
Setup asks for your domain only, and there is no screen for adding tracked competitors yet. Your plan’s competitor allowance is enforced, but there is no way to populate it in the product today.
rekkal does not yet discover and add rivals it finds in answers on its own.