Skip to content
Georadar
Sign inCreate an account

The map of the ground to cover

Noting that you are missing on a prompt does not yet tell you what to do, whereas knowing which domains the assistant consults to answer it points precisely at the ground to stand on.

100%
of cited sources extracted and aggregated
1
marker on your own domains
3
levels of reading: global, category, prompt

Input

The detail behind a citation

When an engine cites its sources, each one is extracted, reduced to its domain and attached to the prompt and the engine that produced it.

  • The domain citedThe URL is normalised to a domain name, so twenty pages of the same outlet count as one player.
  • The source enginePerplexity and ChatGPT in search mode cite abundantly, answers without browsing far less: the detail stays separate.
  • The prompt involvedEach citation stays attached to the question that produced it, so you can drill down case by case.
  • Your own domainsDomains you declare as yours are flagged with a marker, brand site and network sites alike.
sources citées · catégorie café en grain
cuisinecool.fr Vous
34
media-conso-a.fr
28
comparatif-b.fr
21
votre-marque.fr Vous
9
forum-c.com
7

Nombre de citations relevées sur la période, par domaine. Vos propres domaines portent un marqueur.

Output

The answer to “where must we appear”

The source ranking answers the question everyone asks once the diagnosis is in: where do we need to appear to weigh on the answer.

01

The domains that weigh

Ranked by number of citations across your prompt plan, all engines together then engine by engine.

02

Your actual presence

The share of citations pointing to your own domains, which measures what your site really weighs in the answers.

03

Editorial targets

The thematic outlets recurring most often, which indicate the kind of content assistants pick up.

04

The effect of publishing

A network domain appearing among the cited sources after publication reads directly in this ranking.

How it runs

How the list becomes an editorial plan

  1. 01
    Record

    Sources are extracted from every answer that cites them, then aggregated by domain across the whole period.

  2. 02
    Compare

    The ranking is set against your share of voice: a lost prompt with identified editorial sources is a workable prompt.

  3. 03
    Target

    The network verticals matching the cited domains are selected to host the content.

  4. 04
    Verify

    At the next reading, your domains appearing among the sources confirms the content was read by the engines.

Go further

Asking once proves nothing Each customer question is run several times across four assistants, then analysed for brands, rank, recommendation and tone. Explore The lever measurement alone does not give Full articles and in-context inserts, produced from the platform and published on the network sites we run. Explore A crawl today, a citation tomorrow An in-house pixel separates declared assistant crawlers from human traffic and timestamps every visit to your pages. Explore

All features

Your own figures beat any demonstration

Create your free account, approve your prompt plan and launch the first reading. What this page describes then appears on your own brand.

Create my free account Book a demo