How to set up GA4 alerts for conversion drops (2026 guide)

This is the playbook. We will walk through the native GA4 path, then the Looker Studio workaround, then the multi-site alternative. Pick the one that fits the scale you actually operate at.

Method 1: GA4 Custom Insights (built-in)

Custom Insights are GA4's built-in anomaly and threshold detection system. They evaluate metrics on a schedule and send email when a condition is met. This is the right path if you manage one or two sites.

Step-by-step

1. In GA4, click Reports in the left nav.
2. On the Reports snapshot, find Insights at the top right and click View all insights.
3. Click Create, then Custom Insight.
4. Give it a name like "Conversions dropped 30%".
5. Set the Evaluation frequency. Daily is the common pick.
6. Under Segment, choose All Users unless you want to alert on a specific audience.
7. Under Metric, choose Conversions or your specific conversion event (e.g. generate_lead, contact_form_submit).
8. Under Condition, choose % change, set the comparison to Previous period, and set the threshold to -30% or whatever your tolerance is.
9. Add the email recipient (your team alias).
10. Save.

What to expect from native GA4 alerts

Custom Insights work but they have limits:

• Per-property setup. Each client site needs its own Insights. Onboard a new client and you redo the configuration.
• Daily evaluation cadence. GA4 evaluates Custom Insights on a daily schedule by default. A funnel that breaks at 9am could go undetected until the next morning.
• Email-only delivery. No Slack, no SMS. Email is fine but multi-channel matters for severe incidents.
• Hard to triage across sites. Twenty separate alert emails into one inbox makes it hard to spot patterns ("three sites broke at the same time = upstream issue").

Method 2: Looker Studio + scheduled email reports

A common workaround is to build a Looker Studio report that highlights conversion drops across multiple sites, then schedule daily PDF emails. This works but it requires you to build the dashboard, maintain it, and accept that the "alert" is really a daily report.

Tradeoffs:

• You design the layout yourself, so you can build something that supports triage well.
• You can use blended data sources to combine multiple GA4 properties on one page.
• It is a report, not an alert. You only see drops when you open the daily PDF. There is no push notification for "this just broke".
• It requires maintenance every time GA4 changes a field or a client adds a new conversion goal.

Method 3: A multi-site monitor (Parafunnel)

For an agency monitoring 5+ client sites, the per-property setup math stops working. A purpose-built multi-site monitor like Parafunnel solves the same problem with no per-client configuration.

How it works:

1. Connect your Google account via one OAuth flow.
2. Parafunnel auto-discovers every GA4 property your account has access to.
3. Set the alert threshold once at the account level (e.g. "alert when any client's conversion rate drops 30% versus the previous 7 days").
4. Pick delivery channels: email, SMS, Slack.
5. Done. New clients you grant access to are auto-monitored.

Choosing thresholds that do not cry wolf

The fastest way to make alerting useless is to set the threshold too tight. You get alerts every time traffic dips for normal reasons (weekends, holidays, ad campaign cycles) and you start ignoring them. Three rules:

• Compare like-for-like periods. Compare Tuesday to last Tuesday, not Tuesday to Monday. GA4 Custom Insights "previous period" comparison can be week-over-week for stability.
• Use a 25-30% drop threshold. Smaller thresholds trigger on noise. Larger thresholds miss real failures.
• Set a minimum volume. If a small site averages 3 conversions per day, a "50% drop" might be 3 to 1. Skip alerts below a meaningful base rate.

What to monitor across every client funnel

Three layers of alerting cover most agency failure modes:

Layer 1: Catastrophic drops

A 50%+ drop in conversions versus the trailing 7-day average. This catches "the form broke" or "the site is down" failures. High priority, multi-channel delivery.

Layer 2: Sustained underperformance

A 25% drop sustained over 3 to 7 days. This catches "the new design hurt conversion" or "the ad campaign stopped working" failures. Medium priority, daily email digest.

Layer 3: Zero events

The conversion event recorded zero events for 24 to 48 hours when historical baseline shows activity. This catches "the tag was removed" failures, which a percentage-based alert misses if the prior period was already low.

Common mistakes

• Alerting on raw users or pageviews. These fluctuate too much. Conversions and conversion rate are the right metrics.
• Same threshold for every client. A 10-conversion-per-day client and a 1,000-conversion-per-day client have different signal-to-noise ratios. Set the threshold per client tier if possible.
• No escalation path. An alert that goes only to email gets missed during weekends. Severe alerts need SMS or Slack on-call rotation.
• No "alert ack" workflow. When an alert fires, who investigates and what is the playbook? Define this once, save it in a runbook.

Related reading

• How to monitor 20 client funnels without opening 20 tabs
• The one GA4 alert every agency should set up today
• GA4 for agencies: what to track across every client site

Frequently asked questions

Does GA4 send alerts natively?

Yes, via Custom Insights. Threshold-based email alerts per property. Works for one or two sites. The per-property setup gets repetitive at agency scale.

Why are GA4 native alerts not enough for agencies?

Per-property configuration, daily evaluation cadence, email-only delivery, and hard-to-triage across sites. Each new client = redo the Insights config.

What is the fastest way to alert on conversion drops across many client sites?

A multi-site monitor like Parafunnel. One OAuth flow, auto-discovery, one global threshold, email + SMS + Slack delivery, alerts within 1 hour.