- GA4 becomes a growth engine once it knows margin and customer value, not just revenue. Only then can bid strategies optimize for profit.
- Predictive audiences like “likely churning users” only become available once at least 1,000 returning users have met the condition within 28 days, and at least as many have not.
- Since November 2023, GA4 has offered only three attribution models. The data-driven model reveals the contribution of early touchpoints that get zero credit under last click.
- Consent Mode v2, behavioral modeling and first-party measurement through your own domain close data gaps. In GA4, offline purchases can only be attributed via client ID or user ID.
Many companies use Google Analytics 4 as a reporting tool: how many visitors came, which pages they viewed, how much revenue the website generated. That tells you what happened, but not where your next budget should go.
GA4 can do more. Set up properly, it feeds Google's bid strategies the signals they optimize for, shows what each channel contributes and closes the gaps left by cookie banners and browsers. In this article, we cover seven levers that make GA4 a foundation for growth, each with a short how-to.
Data strength starts when GA4 knows margin and customer value
An example: an online store sells premium coffee machines. If it only tracks that a purchase happened, no bid strategy can tell whether the buyer grabbed a one-time deal or will become a loyal customer who keeps reordering accessories and beans.
If the store also passes margin and customer lifetime value (CLV) from its CRM, the goal of an automated campaign like Performance Max changes. It no longer looks for just any buyer, but for customers with high margins and a high likelihood of buying again. Google calls this data strength. We explain how to reflect these values in Google Ads conversion tracking as well in our article on Google Ads conversion tracking.
How to implement it
Send the margin as a custom parameter (for example profit) with the purchaseevent. Upload historical customer values via Admin > Data import as user data, matched via your user ID. Use this data to build audiences for Google Ads.
Predictive audiences reach customers before they churn
With subscription models, such as a gym or a software product, customer lifetime determines profit per customer. GA4 uses machine learning to calculate churn probability: the likelihood that a user who was active in the last seven days won't be active in the next seven. It also provides purchase probability and predicted revenue.
These metrics power audiences like “Likely 7-day churning users.” You can share them with Google Ads to reach at-risk customers with a relevant offer while they're still customers.
Predictive metrics aren't available for every property. Google requires that within the last 28 days, at least 1.000 returning users have met the respective condition and at least 1,000 have not. For purchase predictions, the purchaseevent must also send value and currency along with it.
How to implement it
Under Audiences > New audience, go to the Predictive section and choose the template for likely churning users. If Google Ads is linked to GA4, the audience is automatically shared as a segment.
Data-driven attribution shows which channel kicked off the sale
A typical path to purchase: a customer sees an Instagram story in the morning, clicks a search ad at lunch and buys in the evening through a newsletter with a discount code. Under the last click model, the newsletter gets credit for the entire sale. Make decisions based on that, and you'll conclude Instagram doesn't work and switch off the very channel that created the demand.
The data-driven model distributes credit across all touchpoints based on your property's data. Since November 2023, GA4 has offered only three models: data-driven, paid and organic last click, and Google paid channels last click. First click, linear, time decay and position-based are gone.
Still, attribution is an allocation of credit, not proof. Only an incrementality test, where a control group doesn't see the ads, shows whether a channel actually drives additional sales. For major budget decisions, we recommend both.
How to implement it
Under Admin > Data display > Events > Attribution settings, set the reporting model to Data-driven. In the Advertising section, compare models under Attribution models, and see which touchpoints your customers came through under Attribution paths.
Measuring through your own domain makes data more reliable
Privacy sounds like less data. In practice, it can lead to better data. Ad blockers and browser restrictions like Safari's Intelligent Tracking Prevention (ITP) often block scripts loaded from Google's servers. Those visits and purchases then never show up in GA4.
Two solutions address this:
- With Google tag gateway, Google tags load through your own domain. Setup is relatively quick; you'll find the details in our article on Google tag gateway.
- With server-side tagging via your own server container in Google Tag Manager, data goes to your server first and only then to Google. You control what information gets passed on, and your website loads fewer scripts in the browser.
Combining both is the most robust approach. Neither changes anything about consent: what the user declines stays declined.
How to implement it
Set up a server-side container in Google Tag Manager. The web container sends the data to the server container, where a GA4 tag forwards the cleaned data to your property.
With gross profit as the target, your campaigns optimize for margin
For a long time, ROAS (return on ad spend) was considered the most important metric. But revenue is not profit. Example: at a €1,000 sales price and a 5% margin, €50 remains. For beans at €50 and a 60% margin, it is €30. A bidding strategy that optimizes for revenue favors the machine: it brings twenty times the revenue but only €50 in margin. Two bags of beans already bring €60.
The solution is optimizing for POAS (profit on ad spend). Instead of gross revenue, you pass gross profit: revenue minus cost of goods sold (COGS). Ad spend is the denominator. The bid strategy then learns which sales are actually worth it. How much difference this makes in a real account only becomes clear in comparison:
How to implement it
Create a custom metric for gross profit in GA4 and send it with every transaction. Import this conversion into Google Ads and make it the primary basis for your bid strategy. Cost of goods should come from your ERP system, not from an estimate.
Behavioral modeling closes the gap left by declined cookies
Many users decline analytics in the cookie banner. Without further measures, those visits are invisible to GA4: campaigns look weaker than they are and may be cut for the wrong reasons.
GA4 can estimate this gap using behavioral modeling. The model compares the behavior of users who consented with the cookieless signals from those who declined, and fills the reports with modeled data. It requires Consent Mode v2 and a minimum data volume, which smaller properties often don't reach.
How to implement it
Implement Consent Mode v2 in “Advanced” mode, if your data protection officer approves. Then, under Reporting identity in GA4 Admin, select Blended. We explain which Consent Mode setup is right for you in our article on Google Ads conversion tracking.
Offline data shows what your online ads drive in store
If you also run retail stores, showrooms or partner dealers, you know the ROPO effect (research online, purchase offline): a customer spends weeks researching espresso machines on your website, then buys in the store. That sale never shows up in your online reports, and the advertising that triggered it looks weaker than it is.
GA4 can attribute offline purchases to earlier online sessions via event data import. But this only works if the client ID or a user ID is available at the time of purchase, for example because the customer uses an account or loyalty card linked to the website. GA4 doesn't match hashed email addresses from your POS system. For that, Google Ads and Data Manager are the right route.
This lets you prove your ads' impact on in-store revenue as well, so your budget is measured by its full contribution to profit.
How to implement it
For online contacts, store the GA4 client ID or your user ID in your CRM so it's available when the purchase happens later. Upload in-store sales with this identifier, timestamp and value as a CSV via Admin > Data import.
Conclusion: GA4 delivers when it knows your profit
All seven levers share one foundation: GA4 can only guide your decisions as well as the data you give it. Complete measurement through your own domain, Consent Mode and modeling make sure the data is accurate. Margin, customer value and offline purchases make sure it measures the right things.
That's why our rule is: measure first, then advertise. Report only revenue to GA4, and you get campaigns that optimize for revenue. Report profit, and you give your budget a goal that shows up on the bottom line.
