The first 90 days of measurement work
This is the order the measurement work happens in, and why. It describes process. No results are promised on these dates.
Week 1: agree on what counts
We list every conversion the company reports and check how credit is assigned. Google Analytics offers three models in its attribution settings: data-driven, paid and organic last click, and Google paid channels last click. Data-driven distributes credit using your own account's data. The last-click models give 100% of the value to the final channel clicked. Two teams reading different models will report different numbers from the same data, so we pick one and write it down.
Month 1: fix collection
Consent comes first. Google's consent mode passes each visitor's cookie choice from your banner to Google tags. It does not supply the banner. When analytics storage is denied, no Analytics cookies are set or read, and Google models the gaps from cookieless pings. Legal and marketing should both know that some reported conversions are modeled. Tags, CRM fields and naming rules get cleaned up in the same pass.
Month 3: own the raw data
By now the BigQuery export is running. It sends raw, unsampled Analytics events to a data warehouse once a day, with an optional streaming export that Google describes as best effort. Standard properties have a daily export limit of 1 million events. With raw events beside CRM revenue, finance can check marketing's numbers without waiting on a platform report.
From here the work moves to quarterly planning and to the internal tools your team has been waiting on, built as custom business software. For background, read what marketing attribution is.
