> ## Documentation Index
> Fetch the complete documentation index at: https://docs.synscribe.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Attribution for AI Traffic and Pillar Pages

> We attribute to business outcomes at the pillar level and to AI-answer-engine traffic, never to rankings or impressions as ends in themselves.

A performance review answers one question: did *this pillar* produce enquiries, sign-ups, or paid
conversions? We answer it by reconciling **GSC** (what ranks) against **PostHog** (what converts)
at the pillar level, with AI-answer-engine traffic tracked as its own channel. Rankings,
impressions, and raw traffic are leading signals, never the goal in themselves.

## Rankings are a leading indicator, not the goal

Step 8 of the loop is *"validate or invalidate hypothesis,"* so attribution exists to help you
decide which hypotheses to double down on next month. The join is what makes
it actionable. GSC tells you what got seen and clicked. PostHog tells you what those visitors
actually did. Put the two together and you can see which pillar or intent drives revenue, which is
exactly what you scale next cycle.

Tracking AI-referred traffic separately matters because that surface is the growth edge
([1.2](/theory/two-search-surfaces)) and is invisible to classic rank tracking. A page can be
cited heavily by ChatGPT while barely moving in Google's blue links.

## Conversion, staged down the funnel, per surface

**Wonderchat's** declared key metrics are the template, and notice there's no "traffic" anywhere
in the list:

> **"Enterprise enquiries (form submissions) · Free trial sign ups (landing page → account
> creation) · Subsequent conversion to paid services."**

Attribution is designed into the page rather than bolted on afterward: the reference-repo template
ships the primitives, a *"user journey trace in contact form"* and a *"dual contact form (contact
& get free consultation)."*

## How we think about attribution

Four rules keep the measurement honest. Google organic and LLM referral are separate channels,
measured separately, because most tools miss the AI channel entirely; we match referrers from
ChatGPT, Perplexity, Claude, and Gemini so we can see which pages earn AI traffic. Everything is
tracked at page-and-term level, so every lead traces back to the exact page and search term that
produced it. We instrument the real conversion event, even when it lives on a third-party form or
a redirect, rather than stopping at a pageview and calling it a win. And we refuse vanity metrics:
the number that matters is honest MQL attribution, not impressions.

> ❓ \[needs Raymond: confirm this is the attribution philosophy you want stated as canon]

## Operationalize it

* Set up the dashboard (AI-traffic insight + conversion funnels): [2.6.1 — PostHog dashboard setup](/platform/posthog-dashboard-setup).
* Wire client attribution: [2.6.2 — Set up client attribution](/platform/attribution-setup).
* Reconcile GSC + PostHog into a review: [2.6.3 — Reconcile GSC + PostHog](/platform/performance-review).
* Turn it into client-facing material: [3.7 — Preparing client-review material](/sops/client-review-prep).