How to Use GA4 Custom Alerts for Monitoring Performance & Issues

· 11 min · Google Analytics

GA4 can warn you when performance shifts—before revenue or leads take a hit. Learn how to build practical custom alerts, set thresholds, and respond with confidence.

Why GA4 custom alerts matter (and what they can and can’t do) GA4 is excellent at collecting data, but most teams don’t have time to watch dashboards every day. Custom alerts (often implemented via GA4 Insights, Explorations, and external automation) help you spot problems and opportunities quickly—like a sudden drop in conversions, a spike in paid traffic costs, or tracking that silently broke.

Before setting anything up, it’s important to understand the landscape:

• GA4 includes Insights (some automated, some configurable) that can surface unusual changes. • GA4 can’t natively replicate every “Custom Alert” feature you might remember from Universal Analytics (UA), especially email alerts for any arbitrary metric in standard reports. • The most reliable “custom alerting” in GA4 typically combines: - GA4 Insights (in-product notifications) - Explorations for precise monitoring views - Looker Studio + scheduled delivery - Google Analytics Data API + Apps Script/Cloud Functions for true threshold-based alerts

What GA4 alerting is best for • Detecting tracking issues (e.g., purchases drop to zero) • Monitoring channel performance (e.g., Paid Search sessions spike) • Watching conversion rate changes after releases • Noticing unusual geography/device shifts (potential bot traffic or site issues)

What GA4 alerting is not great for (without external help) • Complex multi-condition alert logic (e.g., “Alert only if conversions drop AND spend stays flat”) • Guaranteed real-time alerting (GA4 data latency can be hours) • Deep root-cause analysis (alerts tell you something changed; you still need a workflow to diagnose)

A practical goal: use alerts to reduce “time to detection.” If it currently takes your team 2–5 days to notice a conversion drop, a good alert system can cut that to same-day.

Choose the right metrics, thresholds, and realistic benchmarks Most alert systems fail because they monitor too much or pick thresholds that fire constantly. Start with a small set of business-critical metrics and add coverage over time.

Pick a monitoring tier: business, marketing, and tracking health A simple model is three tiers:

• Business outcomes - Purchases, revenue, leads (key events) - Purchase conversion rate / lead conversion rate - Average order value (AOV) • Marketing performance - Sessions and users by channel - Cost metrics (if linked to Google Ads) - ROAS / CPA trends (often better monitored in ad platforms, but GA4 can help) • Tracking health - Events volume (e.g., page_view, purchase) - Checkout funnel progression - Sudden spikes in “(not set)” or “Unassigned” traffic

Use thresholds that match your traffic volume Thresholds should depend on baseline volume. A 20% drop in conversions is meaningful for a site with 200 conversions/day, but may be noise for a site with 5 conversions/day.

Realistic starting benchmarks (adjust to your business):

• Ecommerce purchases - Alert if purchases drop 30–50% day-over-day (DoD) for small stores (<20/day) - Alert if purchases drop 15–25% DoD for medium stores (20–200/day) - Alert if purchases drop 10–15% DoD for large stores (>200/day) • Lead generation (forms/calls) - Alert if leads drop 25–40% DoD (common volatility) - Alert if lead conversion rate drops 20% week-over-week (WoW) • Traffic anomalies - Alert if sessions spike 50–100% DoD (possible bots, bad tagging, or PR) - Alert if Paid Search sessions drop 20–30% DoD (budget caps, disapprovals) • Tracking health - Alert if purchases = 0 for 6–12 hours during normal selling periods - Alert if page_view volume drops 30–50% (tag not firing, consent changes)

Prefer percentage change + minimum volume A strong pattern is: “percentage change AND minimum baseline volume.” For example:

• Alert when purchases drop >25% compared to yesterday AND yesterday had ≥40 purchases.

This avoids false alarms on low-volume days.

Set up GA4 Insights for in-product custom monitoring GA4’s Insights are the closest built-in replacement for UA-style alerts. They can surface anomalies and, in some cases, be configured around specific conditions.

Where to find Insights In GA4, look for Insights in:

• The Home workspace (cards and insights) • Reports snapshot (insight cards)

You’ll typically see:

• Automated insights (GA4 detects unusual changes) • Suggested insights (based on your usage)

How to configure an insight (practical workflow) Exact UI labels can change, but the workflow is consistent: define what to monitor, the comparison window, and the threshold.

Go to Home (or Reports snapshot) and open Insights. Choose Create (or Create new insight) if available. Select the metric (e.g., Purchases, Total revenue, Key events). Choose the dimension filter if needed (e.g., Session default channel group = Paid Search). Define the condition (e.g., “decreases by more than 20%”). Set the time comparison (e.g., previous day, previous 7 days). Name it clearly, like “Alert: Purchases down >20% DoD”.

E…