To set up a PostHog Alert for your website activity, log in to PostHog and follow the steps below.

Step 1: Creating an Insight in PostHog

  • In the left sidebar, navigate to the Product analytics tab.
  • Click the New insight button.
PostHog Alerts step1-01
Creating a New Insight in PostHog from the Product Analytics Tab
  • A new page will appear. Fill in the required fields based on the metrics or events you want to track.
PostHog Alerts step1-02
Setting Up and Saving a Pageview Insight in PostHog

Once you’ve finished setting it up, click the Save button.

Step 2: Creating a PostHog Alert for the Insight

After saving the Insight, click the Alerts button. In the pop-up that appears, click New Alert.

PostHog Alerts step2-01
Creating an Alert for the Saved Insight in PostHog
  • Define the conditions that will trigger the alert.
  • In the Notify field, select the team members who should receive the notification.
PostHog Alerts step2-02
Configuring Alert Conditions and Creating the Alert in PostHog

Click Create alert to finalize.

Note: Selected team members will receive alert notifications via email.

You can read the PostHog guide on this topic here to fully understand it.

Decorative

Did you know?

We can just do things for you!

Contact Us

learn more:

Discover how to run A/B tests on WordPress and Shopify to optimize your website performance.

Frequently asked qusetions

Yes. Alerts are always created from an insight. You either:

  1. Create a new insight (e.g. a trend of $pageview on your main website), and

  2. Click the Alerts button on that insight to add an alert.

This is exactly what your article walks through: first build the insight you care about, then attach an alert to it.

At the time of writing, alerts are supported on all Trend insights. (+)

You can’t yet put alerts directly on funnels, retention, or paths; instead, you create a trend that represents the metric you care about (pageviews, events, formula-based metrics like conversion rate) and attach the alert to that trend.

PostHog supports two main alert types on a series in your trend insight:

  • “Has value” – checks if the insight value is more than or less than a number you set (e.g. pageviews > 5,000).

  • “Increases by / decreases by” – checks if the metric changed by a certain amount or percentage vs. the previous period (e.g. conversion rate drops by 10%).

You pick:

  • The series (which metric in the graph to watch),

  • The type (absolute or relative),

  • The threshold (number or percentage).

This gives you enough flexibility to monitor both hard thresholds (e.g. “404s above 100”) and changes (“signups down 20% week over week”).

Alerts are evaluated on a schedule, not in true real-time. When you configure an alert, you choose how often PostHog checks it: hourly, daily, weekly, or monthly.

That means:

  • You’ll get notified after the evaluation window, not the instant an event happens.

  • For “critical” website metrics (e.g. traffic collapse, spike in 404s), hourly checks are usually a good balance between speed and noise.

Yes. When an alert fires, PostHog can:

  • Send email notifications to the users you select, and

  • Trigger CDP destinations, including Slack and other integrations. (+)

The flow is:

  1. Create and save your alert.

  2. Add a CDP notification to that alert.

  3. Choose a template (e.g. Slack), configure the destination (Slack connection + channel), and enable it.

This lets you route important website alerts into Slack, SMS, or other channels used by your on-call or growth teams.

Yes. As long as you can represent it as a trend insight, you can attach an alert to it. Common website examples:

  • Pageviews where Path name = "/pricing" or /checkout.

  • Visits to your 404 page as a proxy for broken links.

  • Events like signup_completed, checkout_completed, or form_submitted filtered by country, device, campaign, or any event/property breakdown.

You build a trend for that slice of traffic or behavior and then set an alert on that insight. You can also use breakdowns (e.g. referrer, channel, UTM) and have alerts fire when any breakdown value crosses your threshold.

Yes. This is one of the best use cases. You can:

  1. Create a trend insight that calculates a conversion rate using formula mode (e.g. signups ÷ unique visits to a landing page).

  2. Attach an alert to that series to fire when the conversion rate increases or decreases by a certain percentage, like 10%.

That lets you automatically catch:

  • Drops caused by a broken form, a bad A/B test variant, or tracking issues.

  • Positive jumps after a successful experiment or UX improvement.

For A/B tests / experiments, you can monitor top-level metrics this way while using PostHog’s experiments module for statistical significance.

Yes. You can do this by setting an alert on a low or zero value. For example:

  • Create a trend that counts signup_completed per day.

  • Configure an alert that fires when the value is less than 1.

This pattern works for:

  • No signups, no purchases, or no pageviews.

  • Critical transactional events (e.g. no billing emails sent, no webhook events received in a day).

It’s essentially “absence detection” using a has value alert with a low threshold.

You can manage alerts in two places:

  • From the insight:

    • Click the Alerts button on the insight.

    • You’ll see all alerts for that insight, where you can edit or delete them.

  • From the Alerts tab:

    • Go to Product analytics → Alerts in the left sidebar.

    • This shows a list of all alerts across your project so you can quickly review, adjust thresholds, or remove noisy alerts.

Yes. PostHog’s “pay-as-you-go” model includes all features on the free tier; you pay only once you pass the free usage allowance (for analytics events, replays, feature flags, etc.).

In practice, that means:

  • You can use alerts on your website insights even on the free plan.

  • The main limit is how many events you send, not whether alerts are enabled.

They solve different problems:

  • Alerts = notify you when a metric breaches a threshold or changes by a certain amount (e.g. “404s increased by 50%” or “signups dropped below 50 per day”). They’re condition-based and run on a schedule.

  • Subscriptions = send regular snapshots of an insight or dashboard to email or Slack on a fixed cadence (e.g. weekly traffic report), regardless of whether anything unusual happened.

For website monitoring, you’ll usually:

  • Use alerts for “something is wrong or interesting, wake me up.”

  • Use subscriptions for routine reporting to the team.

Related blog posts

Leave a Reply

Your email address will not be published. Required fields are marked *