How do you measure the effectiveness of your marketing campaigns with GA4?

A marketing campaign (an email send, a social ad, a partner operation, a product launch) costs budget and time. Knowing whether it worked means measuring it on what it was meant to produce, not on what it displayed. GA4 provides everything you need, provided you prepared the measurement before launch. Here is the method in five steps.

Step 1: set the objective and the metric before you launch

A campaign with no numerical objective is impossible to evaluate. Phrase it as "X conversions of type Y at under €Z each": 50 quote requests at under €40, 200 sales with a return on ad spend of 4, 300 webinar sign-ups. The matching metric has to be a key event that already exists in GA4, with a value. If it doesn't, create it and test it before the campaign, not after.

Step 2: tag every link

GA4 can only isolate a campaign if its links carry consistent UTM parameters: source, medium and a unique campaign name (lancement-produit-x-2026-09). Google Ads uses auto-tagging. The convention is described in tagging your digital campaigns. A forgotten link is a conversion lost to the campaign and won by "direct".

Step 3: monitor during the campaign

In Reports > Acquisition > Traffic acquisition, with the "Session campaign" dimension, find your campaign. Three columns to watch from the first days:

• Sessions : is the campaign bringing traffic? Zero volume after 48 hours signals a tagging or delivery problem.
• Engagement rate : are visitors sticking around? Below 40%, the landing page isn't living up to the promise.
• Key events : is the campaign producing the result you're after? Compare the pace with the volume you expected.

These early checks let you adjust (page, message, targeting) instead of noting the failure at the end. For a paid campaign, add cost per conversion, visible in GA4 for Google Ads and to be worked out by hand for the other platforms.

Step 4: work out the result at the end

Three calculations, in this order:

1. Cost per conversion = total campaign cost (media + creative + time) / key events attributed to the campaign.
2. Return on investment = (value of attributed conversions − cost) / cost. A €3,000 campaign generating 60 quote requests worth €150 each (€9,000) has an ROI of 200%. See how to calculate marketing ROI.
3. Assisted conversions : in Advertising > Attribution > Model comparison, the difference between the data-driven model and last click. A social campaign may show 20 conversions on last click and 45 data-driven; the second figure reflects its contribution better (see attribution in GA4).

Step 5: isolate the real effect with a test

Attribution shares the credit between measured channels, but it doesn't prove causation: a campaign can "convert" customers who would have bought anyway. Two tests within reach of a small business:

• The control group : for an email send, leave 10% of your list out, then compare the purchase rate of the two groups over two weeks. The difference is the email's real effect.
• The geographic test : run the campaign in one area and not in a comparable one, then compare how total conversions move.

These methods are particularly useful for awareness campaigns and for SMA, whose effect often travels through other channels, as explained in GA4 and SMA.

The mistakes that distort the verdict

• Judging on impressions and click-through rate : those are delivery metrics, not result metrics.
• Comparing campaigns run over different periods without allowing for seasonality.
• Ignoring consent : 20 to 40% of visits aren't measured; the real result is higher by that much, for every campaign.
• Stopping measurement on the closing day : conversions from a B2B campaign keep arriving for weeks. Set your attribution window accordingly.
• Adding up conversions from several sources (GA4, Meta, Google Ads): each tool attributes the same sales to itself.

So you don't redo these calculations by hand for every campaign, bring costs and conversions together in a marketing dashboard that reconciles them automatically.