You ship a feature and need to know immediately when adoption stalls or spikes. Amplitude's real-time alerts let you watch feature adoption as it happens, triggering notifications when adoption drops or deviates from baseline. Here's how to set them up so you get alerted before adoption becomes a problem.
Track Feature Adoption Events
First, instrument your feature to send adoption events. This gives Amplitude the data it needs to set up meaningful alerts.
Initialize the Amplitude SDK and track first-use
Import the Amplitude JavaScript SDK and fire an event when users first interact with the feature. Trigger this on page load or when the user clicks the primary CTA—be explicit about what 'adoption' means in your context.
import * as amplitude from '@amplitude/analytics-browser';
amplitude.init('YOUR_API_KEY', {
userId: user.id,
defaultTracking: true
});
// Track feature adoption
amplitude.track('Feature Adopted', {
feature_name: 'new_dashboard',
adoption_date: new Date().toISOString(),
user_plan: user.subscriptionTier
});Set user properties to classify adopters
Use setUserProperties() to mark adopters with a user property. This lets you segment by adoption status and build alerts around specific user cohorts.
amplitude.setUserProperties({
'has_adopted_new_dashboard': true,
'first_adoption_date': new Date().toISOString(),
'adoption_cohort': 'early_adopter'
});
// Track ongoing engagement
amplitude.track('Feature Engagement', {
feature_name: 'new_dashboard',
action: 'view_report',
session_duration: 240 // seconds
});Create a Segment for Adopters
Once adoption events are flowing, create a segment in Amplitude to isolate adopters. This segment becomes the target for your alert conditions.
Build a segment in the dashboard
In Amplitude, go to Analytics > Segmentation and click Create Segment. Name it 'New Dashboard Adopters'. Define the segment as users where the Feature Adopted event occurred AND the has_adopted_new_dashboard property is true within the last 30 days.
Refine the segment with properties
Add filters to target the right audience. For example, only include users on the 'pro' or 'enterprise' plan if that's who you're rolling out to. This prevents alerts from firing on noise from free-tier users.
// Example: Segment definition structure for Amplitude API
// GET /api/2/cohorts to list existing cohorts
const segmentFilter = {
filter: {
operator: 'and',
predicates: [
{
stream: 'events',
name: 'Feature Adopted',
filters: [{
name: 'feature_name',
operator: 'is',
value: 'new_dashboard'
}]
},
{
stream: 'users',
name: 'user_plan',
operator: 'is',
value: ['pro', 'enterprise']
}
]
},
lookbackWindow: 30 // days
};Create and Configure a Real-Time Alert
With your adoption segment ready, set up a real-time alert that notifies you when adoption metrics deviate from normal.
Navigate to Alerts and create a new alert
In Amplitude, click Alerts in the left sidebar. Click Create Alert and select Real-time as the alert type. Choose Trigger when an event or user count drops/rises to monitor adoption velocity.
Set your alert condition
Define the metric: count of Feature Adopted events or active users in your 'Adopters' segment. Set a condition like 'drops below 20 events per hour' or 'active users in segment fall below 50'. Choose a time window (1 hour, 6 hours, 1 day) that matches your expected adoption cadence.
// Example: Real-time alert configuration
const alertConfig = {
name: 'New Dashboard Adoption Drop',
type: 'realtime',
metric: 'count',
event: 'Feature Adopted',
eventFilter: [{
name: 'feature_name',
operator: 'is',
value: 'new_dashboard'
}],
condition: {
operator: 'below',
value: 20 // threshold: fewer than 20 adoptions per hour
},
windowSize: 'hour',
notificationChannels: ['slack', 'email'],
slackWebhook: 'https://hooks.slack.com/services/YOUR/WEBHOOK/URL'
};Set notification channels and save
Select where to send alerts: Slack for immediate visibility, Email for record-keeping, or both. Save the alert. Amplitude will now monitor adoption in real-time and notify you when the condition is breached.
Common Pitfalls
- Tracking adoption too broadly (e.g., 'viewed the feature') instead of a concrete action. Be specific: 'completed first workflow' or 'created a report'—fuzzy definitions lead to noisy alerts.
- Alert thresholds that are too sensitive cause alert fatigue. New features have volatile adoption curves; give them a week of baseline data before tuning alerts.
- Forgetting to segment by user properties. An adoption drop might be normal if it's only affecting a low-priority user segment. Always layer in plan tier, company size, or other cohort filters.
- Not accounting for expected adoption patterns. If your feature rolls out gradually (beta users first), an alert calibrated to the full rollout will fire constantly during early phases. Adjust thresholds as rollout expands.
Wrapping Up
You now have real-time visibility into feature adoption. Monitor how quickly users embrace new features and get alerted before adoption becomes a problem. If you want to track feature adoption automatically across tools and surface insights without manual threshold tuning, Product Analyst can help.