Skip to main content
Help Center
Campaign Specialist 6 min read·Updated Jun 30, 2026

Conversion tracking & Event Match Quality — making your numbers trustworthy

Server-side capture survives ad blockers and iOS, EMQ scores your match quality 0-10, and the EMQ Doctor lists exactly which fields to fix.

Every bidding decision the Campaign Specialist makes is only as good as the conversion data underneath it. If a third of your conversions never reach the ad platform, the bidder is optimizing against a fiction. Fixing tracking is not glamorous, and it's the highest-ROI work in this category.

app.magistry.io/dashboard/tracking
The Tracking page showing server-side conversion dispatch health per platform and Event Match Quality scores
Dispatch health and EMQ at a glance — per platform, on real events.

Server-side capture

Magistry captures events server-side, which makes them resilient to the two big losses: ad blockers, which never see the server, and iOS privacy features, which throttle client-side pixels. The events exist because the order exists — not because a browser script happened to survive long enough to fire.

What EMQ measures

Capturing an event isn't enough — the platform has to match it to a user. EMQ (Event Match Quality) is a 0–10 score of how well your events carry the identifiers that make matching possible: hashed email, phone, name, address, click IDs. A conversion the platform can't match is a conversion your bidding never learns from.

Raising the score

  • Enhanced conversions send hashed customer data (email, phone, address) with each event, recovering matches that identifiers-free events lose. Hashing happens before anything leaves your infrastructure.
  • The EMQ Doctor reads your current events and lists exactly which fields are missing or malformed, per platform — not 'improve your data quality' but 'phone is absent on 61% of purchase events'.
  • Consent state and gclid capture quality are surfaced in Settings, because a missing click ID or a mis-wired consent banner silently zeroes out matching for affected sessions.

Verify with a real order

When the setup looks right, place a real test order and watch it land: correct value, correct attribution, matched on the platform side. One real order through the full path is worth any amount of synthetic validation. Do it once at setup and again after any checkout change.

Bad tracking doesn't look broken. It looks like ads that mysteriously underperform. Check the plumbing before blaming the creative.

// still stuck?

Talk to a human who knows your store.

Live chat Mon–Fri 09:00–18:00 CET, email with a 24-hour SLA, or a 30-minute call. Support reads the same decision_log you do.

Free to read · No email gate · Versioned with the platform