Observation · Microsoft Copilot

How should a new website measure citations in ChatGPT and Perplexity?

Copilot recommended a fixed prompt panel, citation share of voice, AI referral outcomes and separate tracking for mentions, citations and clicks.

15 November 2026, 18:34 CSTen-US (English answer; zh-CN interface)Run 3 of 3Method v0.1
Question

Recorded prompt

“How should a new website measure citations in ChatGPT and Perplexity?”
Product
Microsoft Copilot
Surface
Microsoft Copilot web chat with web search
Model
Copilot Auto (model version not exposed)
Locale
en-US (English answer; zh-CN interface)
Observed answer
Open session ↗
Observed
15 November 2026, 18:34 CST
Fresh conversation
Yes
Answer snapshot

What the answer said

Copilot proposed three measurement layers: citation share of voice across a stable target-query set, AI referral traffic and conversions in analytics, and separate counts for brand mentions, source citations and click-throughs. It recommended a compact monthly dashboard covering citation count, share of voice, prompt coverage, referral sessions, conversion rate, top cited pages and prompt-level results. It warned that AI referral attribution can be incomplete and said traditional rankings or domain authority should not be treated as direct citation metrics.

Linked citations

6 recorded sources

Limitations

What this observation cannot show

Authenticated Microsoft Copilot web-chat run in a fresh conversation. The response proposes a monitoring design and does not establish that the cited third-party benchmarks are independently validated.