If your conversion rate dips, you want to know immediately—not in next week's analytics review. Amplitude's alert system lets you catch significant drops the moment they happen, whether you're tracking signup completion, purchase conversion, or feature adoption. This matters because even a 1% drop can mean thousands in lost revenue if you have volume.
Track Conversion Events Correctly
Before setting up alerts, you need the right events flowing into Amplitude.
Log conversion events with consistent naming
Track your conversion event with the same event name every time. If you're measuring purchase conversion, use a consistent event like purchase_completed. Include relevant properties like product ID, amount, and user segment so you can segment by traffic source, cohort, or product line later.
// Track a conversion event in Amplitude
amplitude.track('purchase_completed', {
product_id: 'prod_12345',
amount: 99.99,
currency: 'USD',
user_segment: 'free_trial',
traffic_source: 'organic'
});Track the baseline denominator event
For Amplitude to calculate conversion rate, you need both the numerator (conversions) and denominator (total qualifying events). Track page_viewed for all visitors to your conversion page, and purchase_completed for those who convert. The alert will calculate the ratio automatically.
// Track page view (denominator)
amplitude.track('page_viewed', {
page_name: 'checkout',
user_type: 'premium'
});
// Track conversion (numerator)
amplitude.track('purchase_completed', {
page_name: 'checkout',
user_type: 'premium',
amount: 99.99
});product_type in one flow and product_category in another, you'll fragment your data and make segmentation harder down the line.Create the Alert in Amplitude
Once your events are flowing, build the alert itself.
Build a conversion rate chart first
Go to Events > Segmentation and create a chart that shows your conversion rate over time. Set the event to purchase_completed, then use Break Down By to segment it by source, product, or user type if you want to alert on specific segments. Save this chart—you'll reference it for the alert.
Create an alert from the chart
Click the three-dot menu on your chart and select Create Alert. Choose the metric (Conversion Rate), and set your threshold. If your baseline is 5%, alert when it drops below 3.5%. Don't set it too tight—you'll get alerts on natural variance.
Configure frequency and time windows
Set how often Amplitude checks the condition—Hourly for high-volume products (1000+ users/hour), Daily for medium volume, or Weekly for low-traffic products. Choose a time window that matches your traffic pattern. For hourly checks, use a 2-4 hour lookback window so you're not alerting on 15-minute blips.
// Conceptual example of Amplitude's internal alert logic
const config = {
metric: 'conversion_rate',
threshold: 0.035, // Alert if below 3.5%
frequency: 'hourly',
lookback_window: '2h', // Check last 2 hours of data
comparison: 'falls_below'
};
// Amplitude runs this check every hour:
// if (conversionRate(lastTwoHours) < 0.035) { notifyTeam(); }Set Up Notifications
Choose your notification channel
When creating the alert, navigate to Notification Settings. Select Slack to post directly to a channel where your team monitors issues, Email for individual notifications, or Webhook if you want to trigger downstream actions (like a Jira ticket or PagerDuty incident).
Connect Slack and test the alert
If using Slack, Amplitude will prompt you to authorize your workspace and choose a channel. Pick the same channel where you post alerts from your other tools—keep incident responses centralized. Send a test alert to verify the message arrives and includes the metric value.
// Your webhook receives a POST payload like this:
{
"alert_name": "Purchase Conversion Rate Drop",
"metric": "conversion_rate",
"current_value": 0.022,
"threshold": 0.035,
"time_window": "2h",
"triggered_at": "2024-03-26T14:32:00Z",
"chart_link": "https://app.amplitude.com/..."
}Common Pitfalls
- Setting your threshold without understanding your baseline variance. If your conversion rate naturally swings between 2% and 6% week-to-week, alerting at 4% generates 50 false positives. Establish a 30-day baseline first.
- Monitoring raw event count instead of the actual conversion rate percentage. Amplitude shows both—make sure your alert is tracking the ratio, not the absolute number of conversions.
- Forgetting to exclude internal traffic or test users. If your QA team runs through checkout 50 times a day, it skews your conversion rate and triggers false alarms every time they test.
- Using hourly checks on a low-traffic product. If you only get 5 conversions per hour, one missing conversion drops your rate from 5% to 0%, triggering an alert on statistical noise rather than a real problem.
Wrapping Up
With alerts configured, you're monitoring conversion rate in real time instead of waiting for a weekly report. You catch drops before they cascade into bigger revenue problems. If you want to track conversion rate alerts across multiple tools—Amplitude, Mixpanel, Stripe, GA4—from one dashboard, Product Analyst can help consolidate them.