All posts Guide

Measuring reach: how Niched attributes LLM and search traffic

A look under the hood at how the tracking snippet classifies referrers into ChatGPT, Perplexity, Gemini, search, and direct — per placement.

Reach you can't measure is reach you can't improve. Here's exactly how Niched turns a page view into an attributed, sourced data point.

One tiny snippet

Each verified site installs a single asynchronous script in its header. On every page load it sends a small beacon — the path, the full URL, the document referrer, and your site's tracking token — to our endpoint. It's non-blocking and adds no measurable load time.

From referrer to source

We classify the referrer host into a human-readable source:

  • LLM answer engines: ChatGPT, Perplexity, Gemini, Claude, Copilot, You.com, Poe.
  • Search engines: Google, Bing, DuckDuckGo, Yahoo.
  • Direct / other: everything else, including no-referrer visits.

Anything from an answer engine is flagged is_llm = true, which is what powers your "AI traffic" numbers.

Attributed to the exact placement

Because the beacon carries the URL, we match each hit to the specific placement living at that address on your site — and validate that the URL's host matches your verified domain before we attribute it. No cross-site spoofing, no inflated counts.

What you get in the report

Per placement: total visits, LLM visits, and a breakdown by source. That's the difference between "I think this helped" and "Perplexity sent 214 readers to this article last month."

Attribution is the whole point. Everything else is vanity.
For creators

Ready to get placed?

List your article with a points bounty and get it published on GSC-verified sites — then watch it surface in ChatGPT, Perplexity & Gemini. Start free.

Get placed
Keep reading
Palette (test only)
Contrast is --brand behind white text. base.html asks for ≥7:1. Hide this with ?palette=0.