When operating a privacy-first engineering education and skill certification platform (such as our **GstarCAD Global Academy** portal with interactive CAD assessments, TCO calculators, and D3.js concept graphs), standard pageview analytics miss the most important signals: which engineering tracks (`BIM`, `Mechanical CAD`, `Civil`, `Drafting`) users complete and where they drop off during multi-question technical exams.
Here is the zero-PII Matomo Tag Manager (MTM) and `_paq` event architecture we use to measure engineering learning progression without third-party tracking cookies:
-–
### 1. Tracking Multi-Track Skill Certification Progression Without Overloading Hits
Instead of firing a network hit on every single multiple-choice option hover, we batch question-level telemetry and fire structured `_paq.push([‘trackEvent’, …])` payloads only at deterministic state transitions (`Lesson Start`, `Checkpoint Saved`, `Track Passed`):
```javascript
window._paq = window._paq || [];
function recordCadTrackMilestone(trackId, lessonId, scorePct) {
window._paq.push([
'trackEvent',
'CAD_Certification_Engine',
\`Track\_${trackId}\_Lesson\_${lessonId}\`,
'Score_Percentage',
Math.round(scorePct)
]);
}
```
-–
### 2. Using Action-Scoped Custom Dimensions for CAD Discipline Segmentation
By configuring two **Action-Scoped Custom Dimensions** in Matomo (`Dimension 1: Engineering_Track` and `Dimension 2: CAD_Experience_Level`), we can correlate documentation reading depth with certification pass rates across all 6 engineering tracks:
* You can see this client-side assessment engine in action on our [GstarCAD Skill Certification & Assessment Portal](https://gstarcademy.com/quiz) and our interactive [AutoCAD to GstarCAD TCO Savings Calculator](https://gstarcademy.com/migration).
For other teams running Matomo on Single-Page Application (SPA) quiz or learning engines, do you prefer firing `trackSiteSearch` or custom events when users filter local knowledge base graphs?