A SaaS company should track five groups of metrics in GA4: acquisition (new users and trial or demo requests by channel), activation (the share of sign-ups who reach a first moment of value), feature adoption, retention (returning users by cohort) and revenue events such as upgrades and cancellations. None of these come out of the box. GA4 needs custom events, a user ID at login and a few user properties, and the revenue side usually has to come from your billing system or CRM.
Below is the setup we use for B2B SaaS clients: which metric maps to which event, what the data layer should contain, and which reports and dashboards each team actually opens.
Which SaaS metrics should you track in GA4?
Start with the questions the business asks, then map each to an event or property. This is the prioritization we work from:
| SaaS metric | GA4 event or property | Effort | Question it answers |
|---|---|---|---|
| Trial or demo requests | sign_up or generate_lead event with a plan or form_type parameter |
Low | Which channels bring trials? |
| Activation rate | A custom activation event, such as first_project_created |
Medium | Do new users reach value? |
| Feature adoption | feature_used with a feature_name parameter |
Medium | Which features stick? |
| Retention | Returning users by cohort, subscription_status user property |
Medium | Who keeps coming back? |
| Upgrades and cancellations | plan_upgraded, subscription_cancelled |
Medium | Where does revenue move? |
| Subscription revenue | purchase with transaction_id and value, or offline import |
High | Which channels bring paying customers? |
Churn rate, MRR and customer lifetime value belong in your billing system. GA4 can explain the behavior behind them, but it should not be the source of those numbers.
How do you set up GA4 for a SaaS product?
- One property, the right streams. A marketing site and a web app on the same root domain can share one web data stream. If the app lives on a different domain, set up cross-domain measurement, or sign-ups will show the marketing site as a referrer.
- A user ID at login. Send your internal account or user ID (never an email address) as
user_id. It links sessions across devices and lets GA4 follow a visitor from first visit to paying customer. - A written event plan. Name the events, their parameters and which ones are key events before anyone builds them.
- Implementation through Google Tag Manager, reading values from the data layer.
- Basic hygiene: filter internal traffic, set event data retention to 14 months, list your payment provider as an unwanted referral.
What should your SaaS data layer contain?
The data layer is the object on the page that GTM reads from. For SaaS, push a few facts about the user and account on every logged-in page:
user_id: your internal ID, not personal data.subscription_status: trial, free, paying or churned.plan: for example Basic, Pro or Enterprise.trial_end_dateor days left in trial, to see where trials stall.account_role: admin, editor or viewer, which matters in B2B where the buyer and the daily user differ.
Register the ones you want to segment by as user-scoped custom dimensions in GA4. GA4 limits how many you can register, so keep the list short.
Which events should a SaaS product send?
Events that mark progress through the customer lifecycle:
- Sign-up:
sign_up(a Google recommended event) when the trial or account is created. - Activation: one event for the action that predicts retention in your product, such as
first_project_createdorintegration_connected. - Feature use: a single
feature_usedevent with afeature_nameparameter, rather than one event per feature. - Subscription changes:
plan_upgraded,plan_downgraded,subscription_cancelled, ideally sent from your backend when billing confirms them.
Mark only sign-up, activation and paid conversion as key events. Those are also the events your ad platforms should learn from; our post on offline conversions via the GA4 Measurement Protocol shows how to send backend events like upgrades.
How do you build SaaS funnels in GA4?
Use a funnel exploration (Explore > Funnel exploration) with your own events as steps. Two funnels cover most SaaS questions:
- Onboarding:
sign_up> profile completed > workspace created > first project > team member invited. - Trial to paid:
sign_up> activation event > pricing page viewed >purchase.
Break the funnel down by first user source to see which channels bring users who activate, not just users who sign up. Making it an open funnel lets users enter at any step, which suits products where people skip onboarding.
How should you segment SaaS users?
By the properties you pushed in the data layer:
- Subscription status: compare what trial users do with what paying customers do.
- Plan: find features that Basic users touch and Pro users rely on, which points to upsell triggers.
- Tenure: first 30 days, 3–6 months, a year or more.
- Feature use: users of a feature against those who never touch it.
- Role: admins versus everyday users in B2B accounts.
What should a SaaS GA4 dashboard show?
One dashboard per team, built in Looker Studio on top of GA4 (and BigQuery once you join CRM data):
- Marketing: sessions, trials and key events by channel and campaign, and cost per trial once ad spend is joined in.
- Product: activation rate, onboarding funnel and feature adoption by plan.
- Customer success: engagement trends for paying accounts and accounts that have gone quiet.
- Leadership: trials, activation and paid conversions over time, next to revenue from billing.
When is GA4 not enough for SaaS?
Once the questions move deep inside the product. GA4 tells you what happened on the website and where visitors came from; product analytics in Mixpanel or PostHog follows individual users and accounts through the app with proper retention tables. Most of our SaaS clients run both, linked by the same user ID. See GA4 vs Mixpanel for how we split the work.
TMT Finance is the clearest example. $4,200 of ad spend had produced two conversions. We integrated product analytics, mapped the application journey and fixed the onboarding friction it revealed before relaunching Google Ads. Cost per lead fell from $2,140 to $48.74, with 90+ data points now tracked across GTM and GA4.
For the full list of tracking faults to check first, see our GA4 masterclass on fixing broken tracking.
FAQ
Should GA4 run inside the SaaS app or only on the marketing site?
Both, if you want to measure activation and feature use. Keep app tracking lean: the events in your plan, a user ID and no personal data in event parameters.
Can GA4 calculate churn and MRR?
Not reliably on its own. GA4 can show cancellation events and users who stop returning, but churn rate and MRR should come from billing. Join the two in BigQuery or Looker Studio when you need both in one view.
Is GA4 good for B2B SaaS with long sales cycles?
For the top of the funnel, yes. For the rest, connect it to the CRM: import closed-won deals as offline conversions, so channels are judged by revenue instead of form fills.
What is the most common GA4 mistake in SaaS?
Tracking everything and marking too many events as key events. Ad platforms then optimize for cheap actions, such as newsletter signups, instead of trials that turn into customers.
Our GA4 consulting covers SaaS event plans, user ID setup and dashboards; get in touch if you want a second pair of eyes on yours.
