Alerts allow you to get automatically notified to any changes in your metrics. This article walks through the steps of creating a new Analytics alert.
There are two types of Alerts available: Analytics alerts and Real time alerts. Analytics alerts monitor performance every 24 hours, whereas Real time alerts monitor performance every 5 minutes (useful for being alerted to critical errors or traffic changes that can have immediate impact).
Creating a new Analytics alert
1. Go to 'Alerts' from the main menu and click 'Set up a new alert'.
2. Select 'Analytics alerts'.
3. Define your alert context, then click 'Next step'.
- Scope: Site, Page group, or Zone.
- Metric: Select one of the metrics listed in the dropdown.
- Device: Choose between all devices, desktop, tablet or mobile.
- Segment: Select from one of of your predefined segments.
4. Choose from one of the following alerts types, followed by 'Next step'.
- 'AI alert'
- 'Manual alert'
5. Configure the alert sensitivity (AI Alerts) and trigger condition (AI or Manual alerts).
6. Choose how you want notifications to be sent: via Email, Slack, or Microsoft teams, then click 'Next step'.
- Slack notifications can be sent to multiple channels, however notifications can only be sent to public channels. See our integration article on Slack for more details.
7. Review the summary, then click 'Create' to confirm.
AI alert configuration
Automatically detects unusual variations over time. Configure the following options:
-
Sensitivity level: Define the level of sensitivity from which the alert will be triggered.
- HIGH SENSITIVITY→ 95% of confidence level
= a strict detection of anomaly (more alerts) - MEDIUM SENSITIVITY→ 99% of confidence level
= a medium tolerance of anomaly - LOW SENSITIVITY→ 99,9% of confidence level
= a relatively high tolerance of anomaly (fewer alerts)
- HIGH SENSITIVITY→ 95% of confidence level
- Trigger condition: Define when your alert notification gets triggered. When the selected metric value goes: 'Above or below', 'Above', or 'Below' the expected range.
How do AI Alerts work?
Thanks to artificial intelligence, Contentsquare is able to predict for any metric - based on the last 30 days' trends until the day of the incident and 2 days before reaching the day of the incident.
The prediction algorithm is based on a combination of linear growth and periodic variation functions.
The next day, we collect the real value of the metric and check whether it falls within the predicted confidence interval. If not, the real value is considered an anomaly that triggers a notification in the notification center of Contentsquare and sends email and in-app notifications (if you choose to be notified these ways).
Manual alert configuration
Triggers upon a threshold defined by you. Choose your threshold type and define the trigger:
- Percentage (%) change limit.
- Fixed limit
Create alerts from Shortcuts
You can create alerts on the fly from Journey Analysis, Page Comparator, Zoning Analysis and Dashboards.
Right-click on the page (Journey Analysis, Page Comparator), zone (Zoning Analysis), or 'more options' icon on a widget (Dashboards) to open the shortcut menu, and click Create an alert.
A summary of your configuration will appear in a side panel. By default, a medium variation AI Alert is proposed. You can edit it and create your alert following the steps above.
Alert ideas
| Alert Type | Metrics & KPIs to monitor |
| Acquisition & Campaigns |
|
| Merchandising & Content |
|
| Design optimizations |
|
| Bug detection & Frustration |
|
Learn more about KPI monitoring with alerts in our article, How to monitor the performance of KPIs.