I moved a Shopify Plus store to server-side GTM last week; Meta event match quality rose from 6.8 to 8.9 and in-platform ROAS ticked up 14% while GA4 revenue stayed within 1%. For those running CAPI server-side, which proxy or vendor is giving you the cleanest dedupe and lowest latency — native sGTM, Stape, or Segment — and how are you validating with BigQuery logs?
On native sGTM, the cleanest dedupe for us was setting purchase event_id to the Shopify order_id and passing it to both the web pixel and CAPI, then exporting Cloud Run logs to BigQuery to query duplicates and p95 latency; that kept GA4 revenue within about 1% and match quality near your 8.9. Stape worked too, but we saw occasional 300ms spikes, so I keep minInstances=1 on Cloud Run. @OP are you also sending _fbp + external_id as Facebook recommends (https://developers.facebook.com/docs/marketing-api/conversions-api/deduplicate-pixel-and-server-events)?
We’ve seen Stape edge out native on p95 latency on Plus by pinning the region close to shoppers and keeping HTTP/2 keep‑alive on; dedupe stayed clean by persisting event_id=order_id in a first‑party cookie and forwarding fbp/fbc to CAPI from the server. For validation, we stream every inbound request from the server container into BigQuery with a source label (web vs capi) and alert on event_id that don’t resolve to exactly two sources or show revenue drift — keeps the pipes unclogged. @jhenry434 have you tried suffixing the event_id with gateway on split‑tender orders?
I can commit 5–6 hours a week and want repeatable labs I can bring Same cadence — MITRE ATT&CK Defender’s Adversary Emulation + SOC Assessments gave us repeatable emulations and detection write-ups; in about 6 weeks we pushed about 10 S
Have you checked out the ATT&CK Defender’s lab? It provides great hands-on experience within 30–60 hours. Might be helpful to track your MTTD improvements too.