First-party research
What AI cites. What Google ranks.
Published measurements from Linkeddit's own work on competitor buying questions, AI citations and Google rankings. Every result has a date and a source.
Published evidence
These are dated observations, rather than a live ranking or a promise of growth.
October 28, 2026 · Original Search Console study
When impressions stay flat but clicks fall
Two equal 28-day windows from our own site: impressions stayed near 100,000 while clicks fell from 662 to 517. Three selected product-relevant queries improved in average position.
Inspect the query selection, daily totals and decisions. This is one site, with no claim that AI Overviews or the latest content changes caused the movement.
Read the study and methodDownload the study data (JSON)October 28, 2026 · Search product evidence
What our own Google checks returned
The Search product page publishes four rows from the first check of our own 47 tracked keywords. Each shows the rank, AI Overview appearance, citation status and leading domain.
None of the four published keywords placed Linkeddit in the top ten. This is an inspectable product output, with no before-and-after improvement claim. The checks cover the first ten Google results in the United States.
Inspect the four results and their limitsSeptember 4, 2026 · Historical Answer Radar baseline
What 15 SaaS buying questions revealed
Our saved summary records 61 attempted samples, 60 successful and 1 failed, across 4 engines. Linkeddit was absent from 9 of the 15 questions in that summary.
This is a historical summary. Raw answers, exact full prompts and per-engine allocations are unavailable. Citation counts are not recommendation rates, and no post-change uplift was measured.
Read the audit and methodDownload the recorded counts and limitations (JSON)Buyer questions and practical methods
What buyers ask about reporting and competitor monitoring
Ten selected public discussions across eight communities, nine with visible replies. A qualitative source register with dates, question themes and commercial-bias notes, not a representative survey.
Download the source register (JSON)Build an AI visibility client report
Question selection, valid-answer denominators, failures and separate business outcomes. Includes a blank evidence log and an explicitly illustrative calculation.
Download the blank answer log (CSV)Research and guides by Om Patel.
How to read these measurements
A source, a mention and a recommendation differ
A citation links to a source. A mention names a brand. A recommendation endorses an option. We identify which observation a published result supports.
Dates and sample sets matter
An engine can answer differently on another day or with different wording. Compare the same questions and disclose failures, engine changes and missing AI Overviews.
A baseline is the starting point
A later result is needed to assess movement. Google clicks, AI recommendations and product outcomes are separate measures; one does not establish the others.
Linkeddit publishes this research and sells the products used to collect it. These examples concern our own work. For our source and correction policies, read the methodology and editorial policy.
Start with your own buying question
The free AI Visibility Checker returns one real Gemini answer. Repeated measurement across six engines is available in Answer Radar.
Free account required. No credit card or trial.