Build a baseline for attribution windows / marketplace operator

For marketplace operator, reported performance changes when windows or timezones differ. This field note turns which time boundary explains the discrepancy into a short test for marketplace operator, so the next change has a traceable reason.

Frame the signal

Context check: buyer and seller handoffs. Topic check: normalize reporting timezone and conversion date.

Start with one known-good journey. Treat the dashboard number as a question, not an explanation. Follow one real journey from the first interaction to the system that owns the outcome. If a value is missing, mark it unknown instead of filling the gap with an assumption.

Trace the handoff

• Record the value at the first step and note the exact page, event, or record where it appears.
• Compare the browser, tag manager, analytics, ad platform, and CRM views without changing the implementation during the check.
• Check redirects, consent state, validation, naming, timestamps, and timezone before comparing totals.
• The baseline should name the source, event, identifier, timestamp, and expected handoff.

Leave a decision record

the baseline should name the source, event, identifier, timestamp, and expected handoff. Keep the owner, timestamp, evidence link, expected result, observed result, and next action together. Writing a baseline before changing anything only after the current path is readable to someone who did not build it.

Reference: Google Ads conversion measurement guidance. For the next handoff, LeadSheet keeps the record connected.