Revenue Attribution
How Attribution Works
The data pipeline: how visitors, referrers, and purchases are connected to build attribution.
Reaudit's attribution engine connects the dots between a visitor arriving on your site and a purchase in Stripe. Here is how the data flows:
Step 1: Visitor Arrives
When someone visits your site, three things can record it: GA4 (via the Google Analytics tag), Cloudflare (via DNS-level analytics), and the Reaudit tracking script. Each records the referrer — where the visitor came from (e.g., google.com, perplexity.ai, direct).
Step 2: Source Normalization
Referrer data from different sources uses different naming conventions (e.g., "google" vs "google.com" vs "Google Organic"). Reaudit normalizes all source names into a canonical form so that data from GA4, Cloudflare, and tracking can be merged without duplicates.
Step 3: AI Detection
The system automatically identifies which visitors came from AI platforms (Perplexity, ChatGPT, Claude, Gemini, etc.) and which page views are from AI bots (GPTBot, ClaudeBot, PerplexityBot, Bingbot AI, etc.). These are separated into AI Referrals and Bot Crawls.
Step 4: Purchase Matching
When a Stripe charge comes in, the attribution engine looks at the customer's journey — their email, the pages they visited, the referrer that brought them, and the time window between their visit and the purchase. It then assigns an attribution confidence:
- High — Direct match: the customer visited from an identifiable source and purchased within the same session or a short window
- Medium — Indirect match: the customer visited from a known source but the purchase happened in a later session
- Low — Weak signal: the purchase could be correlated but the connection is uncertain
- Correlation — Statistical correlation between bot crawls or AI referrals and revenue, without a direct visitor match
Step 5: Aggregation
All data is aggregated by date and by source. The dashboard shows merged referrer data (with GA4 taking priority for visitor counts when available), revenue per source, and time-series data for the chart.
De-duplication
When the same source appears in multiple data streams (e.g., "Perplexity" in both GA4 and Cloudflare), the system uses the highest-priority source for visitor counts (GA4 > Tracking > Cloudflare) while combining revenue data from Stripe.