google-search-console-data skill (period-over-period, striking-distance queries at positions 10–20, cannibalization), pull PostHog with the posthog-analytics skill (per-pillar funnels with AI-referral traffic split out), then reconcile the two into a keep / optimize / double-down verdict per pillar — the input to next cycle’s plan. Attribution (Part 1 §1.10) earns its keep here: it exists to validate or invalidate a hypothesis, so a ranking on its own settles nothing — the join is what tells you which pillar actually drives revenue.
Step 1: Pull GSC (what ranks, what’s slipping, what’s in reach)
google-search-console-data, a compare-to-previous-period sweep of top pages and queries,
a striking-distance pass (positions 10–20 by impressions, the pages worth optimizing rather than
building new), and a query-to-pages check for cannibalization. Lead with the pattern (“clicks up
18%, driven by X, Y, Z”) before the rows.
Step 2: Pull PostHog (what converts, and what came from AI)
posthog-analytics. It reads the Pages-with-AI-Traffic insight and the per-pillar funnels
from 2.6.1, and returns the conversion contribution per pillar
with AI-referred traffic isolated (classic rank tracking can’t see that surface;
§1.2).
Step 3: Reconcile into one picture
🎬 Video planned: reconciling GSC + PostHog into one performance review (pull GSC, pull PostHog, reconcile AI traffic and pillar conversion, then write the review). See the shot-list.Back to: 2.6 Analytics & review · Depends on: 2.6.1 PostHog dashboard setup · 2.1.3 Connect GSC · Theory: 1.10 Attribution