Signal Hub
Tracking, signals and the decisions built on top of them
Not a blog. A connected knowledge system for marketing teams who need to know whether a performance drop is real, or whether the measurement layer moved. Every topic is both an article and a diagnostic step.
Run a free Milestic Signal AuditMeta and CAPI
Meta and CAPI · 4 min
Meta Pixel not tracking purchases
Purchases happen, Meta shows none. Where the event usually disappears.
Meta and CAPI · 4 min
Meta CAPI deduplication
Two paths for the same purchase. How Meta decides they are one event.
Meta and CAPI · 4 min
Event match quality
Weaker matching changes what can be attributed, not what you sold.
Meta and CAPI · 4 min
Purchase event firing twice
Conversions look great, revenue does not match. Usually a duplicate source.
GA4 and GTM
GA4 and GTM · 4 min
GA4 purchase event not firing
Orders exist, GA4 shows nothing. Where the ecommerce event breaks.
GA4 and GTM · 4 min
GA4 revenue not matching Shopify
Some difference is normal. Learn which size of gap is a real problem.
GA4 and GTM · 4 min
GTM duplicate events
One action, two events. The trigger patterns that cause it.
GA4 and GTM · 4 min
GA4 vs Meta discrepancy
They measure different things. Learn when the gap is still a warning.
Diagnostics and decisions
Diagnostics · 4 min
Why did my ROAS suddenly drop?
Six checks that separate a real drop from a measurement drop.
Diagnostics · 4 min
Worse performance after consent update
A consent release can move your numbers without moving your sales.
Diagnostics · 3 min
Can bad tracking affect performance?
Two different questions with two different fixes.
Diagnostics · 4 min
How to know if tracking is broken
The symptom list, and a weekly routine that catches problems early.
How a symptom becomes a diagnostic path
Instead of separate posts, each problem is a node. You start from what you see, and follow the chain until you know whether the signals or the performance moved.
Step 1
Meta ROAS suddenly dropped
Start from the symptom, not from the dashboard you happen to have open.
Step 2
Is the Purchase event firing?
Check whether the event exists at all before you question the campaign.
Step 3
Did Event Match Quality change?
Weaker matching changes what the platform can attribute, not what you sold.
Step 4
Browser events vs server events
Two paths for the same action can mean missing events or double counting.
Step 5
Did consent settings change?
A consent or tracking release can move the numbers on its own.
Step 6
Does GA4 show the same decline?
If only one system moved, the measurement layer is the first suspect.
Not sure whether you have a performance problem or a signal problem?
Your performance may be fine. Your signals aren't.
Run a free Milestic Signal Audit