Case: a timezone mismatch made every campaign look like it died at midnight

Case: a timezone mismatch made every campaign look like it died at midnight

The symptom: dayparting reports showed a hard ROI cliff every day at 00:00, and a spike at the same hour. Buyers were pausing campaigns during 'dead hours' that weren't dead.

Root cause, found in the settings:
— Tracker (Voluum) was set to UTC
— The traffic source reported costs in the account's local time (UTC+2)
— The network posted conversions stamped in UTC-5
— Clicks, costs, and conversions for the same real moment landed in three different hour buckets

The procedure to confirm and fix:
1. Sent one test click and one test conversion, noted the wall-clock time of each
2. Read the timestamp each system recorded for that single event
3. The three stamps were 7 hours apart — proof of the misalignment
4. Standardized the tracker, the source cost import, and the conversion import all to UTC
5. Re-pulled 14 days; the artificial midnight cliff vanished

Outcome: the real dayparting curve appeared. The 'dead hours' buyers had been pausing were actually a profitable block once the buckets aligned — reactivating them added measurable volume.

One timezone, everywhere: tracker, cost import, conversion import. Verify with a single timestamped test event.

Save this SOP. Run this before every launch.