Two Weeks to GA4, and the Longer Game After Cookies

In two weeks, on July 1, Universal Analytics stops processing data. This is the last post I will write about the deadline, I promise, and it comes in two parts: the final checklist for teams cutting it close, and the bigger picture that the GA4 scramble has been obscuring all spring. Because the forced migration was never really the story. It is the opening drill of a decade in which the measurement ground keeps moving, and the teams treating it as a one-time IT project are missing the lesson.

The final fortnight checklist

If your GA4 property has been collecting clean data since winter, this is a victory lap. If not, triage in this order:

• Verify conversions are firing in GA4 and feeding Google Ads directly. Bidding on conversions imported from UA breaks on July 1. This is the item that costs real money if missed, so it goes first.
• Hunt down silent UA dependencies. Looker Studio dashboards, audience lists in ad platforms, automated email reports, that spreadsheet the finance team refreshes. Anything drinking from UA goes dark or stale. Inventory now, migrate or retire deliberately.
• Export your UA history this month. The property will remain viewable for a while after processing stops, but treat access as temporary. Pull the reports you actually revisit into flat files or BigQuery.
• Fix the retention setting and enable BigQuery export if you somehow have not. Two clicks each. Fourteen months and free insurance, respectively.
• Annotate the changeover everywhere. GA4 and UA count differently; sessions and conversions will not match. Write the expected discontinuity down now so nobody diagnoses a July "traffic drop" that is actually a definition change.

The actual story: measurement is becoming modeled

Zoom out from the deadline and a pattern is unmistakable. Apple's tracking changes kneecapped mobile attribution two years ago. Chrome, the last major browser still supporting third-party cookies, has them scheduled for deprecation next year, and while that date has slipped before, the direction has never wavered. Privacy legislation keeps multiplying. GA4 itself, with its modeled conversions and behavioral thresholds, is Google's answer to a world where observing every user directly is no longer on the table.

Put plainly: the era of deterministic, user-level tracking of everyone is ending, and everything replacing it is a model, an aggregate, or an estimate. That is not a tooling change. It is an epistemological one, and it demands different habits:

First-party data becomes the foundation. Consented email relationships, authenticated experiences, purchase history you hold yourself. The brands investing in genuinely valuable reasons for customers to identify themselves are building the only durable measurement asset there is. This is a strategy and creative problem wearing an analytics costume, which is why the growth team keeps crashing brand meetings.

Experiments replace click-trails. When you cannot follow the user, you test the aggregate: geo holdouts, matched markets, incrementality tests. We have written about this before and will keep repeating it, because a team fluent in experiments is immune to most measurement weather.

Marketing mix thinking returns. The old-school statistical approach to budget allocation, unfashionable for fifteen years because click attribution seemed to make it obsolete, is quietly coming back at the exact moment click attribution stops being trustworthy. Worth a place in the 2024 planning conversation for any brand spending seriously across channels.

Treat July 1 as a rehearsal

Here is the reframe for your team retro after the cutover: the organizations that handled GA4 well did not have better analysts, they had better habits. Documented dependencies, owned instrumentation, decisions written down, no single report anyone was afraid to touch. Those habits are exactly what the next disruption will test, whether it is cookie deprecation, the next platform privacy shift, or regulation nobody has drafted yet. Measurement change is now a permanent condition. Build the muscle, not just the migration.