04 · Tracking & Analytics
Know where the money actually goes.
Two million events a day, in production. The point of that volume is not the dashboard — it is being able to answer, precisely, which spend produced revenue and which produced bots.
The problem
Three sources, three answers.
Most teams have GA4, an ad platform and their own database, and all three disagree. Nobody can say which is right, so decisions get made on the number that is easiest to reach.
Underneath that is a measurement problem. Events fire twice or not at all, identity breaks across devices, and a meaningful share of paid clicks were never human in the first place.
Click fraud is the version of this with a direct cost. Budget goes to traffic that cannot convert, and without per-click evidence there is nothing to dispute.
What we build
Measurement you can audit.
- Browser, device and network fingerprinting for identity and risk signals
- Click-fraud detection with real-time scoring and blocking
- GA4 and Google Ads API reporting pipelines
- Custom dashboards over MongoDB event stores
- Cross-device attribution with explicit confidence rather than silent guesses
- Schema.org and JSON-LD automation for structured data coverage
Architecture
From click to conclusion.
- 01collectA small client script plus server-side capture records each event once
- 02fingerprintBrowser, device and network signals produce a stable identifier and a risk score
- 03scoreTraffic is classified in real time; repeat offenders are blocked at the edge
- 04storeEvents land in MongoDB, sized for two million a day with room over
- 05joinAd platform and GA4 data are reconciled against your own order records
- 06reportA dashboard where every number can be traced back to its events
Stack
- Laravel
- JavaScript
- MongoDB
- GA4
- Google Ads API
- Meta API
- ApexCharts
- Schema.org
- JSON-LD
Questions
Asked most often.
Is fingerprinting compliant?
It depends on jurisdiction and purpose, and we treat that as a design input. Fraud prevention and analytics have different bases; we implement consent handling accordingly and document what is collected.
What does the 40% ad-waste figure mean?
It is the average reduction in wasted ad spend across engagements where we deployed fraud detection — spend that previously went to traffic scored as non-human or as repeat offenders.
Does this replace GA4?
No. GA4 stays for what it is good at. We add the layer that reconciles it against your own data so the disagreements become explainable.
Can you handle our event volume?
Two million events a day is the current production figure, on MongoDB. Volume is an architecture question we answer during discovery, not a hard ceiling.
Next step