Anomaly detection in Google Analytics 4 is a statistical technique that Analytics Intelligence uses to identify anomalies in time-series data for a given metric, and anomalies within a segment at the same point of time.
In this article we’re going to take a look at anomaly detection in Google Analytics, and how this new kind of alerting eliminates the need to manually figure out if something was statistically significant.
To understand how anomaly detection works, we need to first examine what an anomaly is.
What is an anomaly?
An anomaly is defined as something that does not conform to the expected. These ‘anomalies’, often referred to in data analysis, are observations that do not conform to an expected pattern.
From a business standpoint, a typical example of an anomaly would be a sudden, unexpected drop in sales or a surge of new users for a brief time period.
Anomaly Detection in GA4
One of the most notable advantages of GA4 is its ability to detect anomalies. With data analysis, collection, and accuracy becoming increasingly important to business success. With Google Analytics 4, you can use Google’s machine learning algorithms to detect anything that might be of value to you.
How does anomaly detection in GA4 work?
Anomaly detection in Google Analytics 4 works as follows:
Firstly, Intelligence selects a period of historic data to train its forecasting model. For the detection of daily anomalies, the training period is set to 90 days. For weekly anomaly detection, the training period is 32 weeks.
Once Intelligence selects a training period, it applies a Bayesian state space-time series model to the historic data in order to forecast the value of the most recently observed data point in the time series.
Last but not least, Intelligence flags the datapoint as an anomaly through the use of a statistical significance test with p-value thresholds based on the amount of data in the reporting view.
How to see anomaly insights in Google Analytics
As we know, Analytics Intelligence regularly scans your data for anomalies. These anomalies are presented as Insights on mobile and web. To see anomaly insights:
- Sign in to Google Analytics
- Navigate to your view
- Use the search field to enter a query, such as:
- “Did anything unusual happen this week?
- “Show me an anomaly in users this month”
Although anomaly detection in Google Analytics is still in its early stages, hopes for it becoming more advanced are ever-present. However, being able to automatically detect anomalies can provide businesses with great insights that they would never have discovered otherwise.