Skip to content
Customer Support

Support Center

Welcome to the Blue Ninja Systems Support Center. Find direct answers to your questions about Authority Infrastructure™, get step-by-step guides, explore learning paths, and submit a ticket if you need hands-on help.

How does attribution confidence scoring work?

Attribution confidence scoring assigns a reliability level to each revenue attribution — high, medium, or low — based on how directly a purchase can be tied to a specific traffic source or campaign.

TL;DR

Attribution confidence scoring assigns a reliability level — high, medium, or low — to each revenue attribution based on how directly a purchase can be tied to a specific traffic source. High confidence means a deterministic, first-party match. Low confidence means a fallback estimate. Performance Infrastructure minimises low-confidence attributions by design through first-party tracking.


Why confidence scoring matters

Most attribution tools give you a number and present it as fact. Performance Infrastructure gives you a number and tells you how much to trust it.

This matters because:

  • Scaling spend based on low-confidence attribution is how brands waste budget.
  • Knowing which attributions are high-confidence lets you make decisions with conviction.
  • Knowing which attributions are low-confidence tells you where to improve your tracking setup.

The three confidence levels

High confidence — Deterministic match

A first-party visitor ID is captured when the user lands on your site and is carried into the checkout metadata when they purchase. The connection between the traffic source and the revenue event is direct and verifiable.

What this requires: First-party tracking installed correctly, visitor ID passed into Stripe checkout metadata.

Medium confidence — Identity linked before purchase

The visitor's identity is established (via email capture, login, or CRM match) before the purchase event, allowing the traffic source to be linked to the revenue event with reasonable certainty.

What this requires: An email capture or login event that links the visitor to a known identity before checkout.

Low confidence — Fallback estimation

The purchase cannot be directly linked to a specific traffic source. A fallback method (last-touch, time-based, or modelled) is used. Performance Infrastructure minimises these by design — but they will always exist for some portion of traffic.

What this means: Treat low-confidence attributions as directional signals, not decision-grade data.


How to improve your confidence score distribution

  1. 1Verify first-party tracking installation — Run the tracking verification check in the Performance Infrastructure portal.
  2. 2Confirm Stripe checkout metadata — Ensure visitor IDs are being passed into Stripe order metadata at checkout.
  3. 3Add email capture before checkout — Even a simple email capture step before the payment page significantly increases medium-confidence attributions.
  4. 4Review CRM connection — If Brevo is connected, verify that contact attributes are being synced correctly so identity matching works.

Common issues

  • All attributions showing as low confidence: First-party tracking is likely not installed or not passing visitor IDs to Stripe. Check the troubleshooting guide.
  • High confidence for some channels but not others: Likely a UTM parameter issue on specific campaigns. Verify UTM consistency across all ad platforms.
  • Confidence dropped after a site change: A site update may have broken the tracking script. Re-verify installation.

People Also Ask

Recommended Next Step