Engagement Level criterion

The Engagement Level criterion is a segmentation criterion that enables defining a .

To learn how to create a segment, please refer to this .

Be careful! This segment criterion is not in real time. The patterns identification of your visitors' behavior will be processed in the hour following the launch of the first campaign using this criterion.

📖 Definition

The Engagement level segment criterion enables you to restrict the visibility of a campaign to visitors based on their engagement on your website.

⚙️ Configuration

This criterion is based on data from the Universal Collect of AB Tasty. According to the behavior of the visitor on the website, AB Tasty is able to sort them into four categories. Visitors who recently visited the website will be evaluated. The association of the VisitorID and the segment will be stored and remain available for actual segmentation when they visit the website again. After the second visit to the website, the visitor can be categorized. These automatically built-in models help you increase the Engagement Level of your visitors by pushing messages or promotions via a campaign.

Heads up ⚡

The category to which the visitor has been assigned is not definitive. It will change over time, depending on their actions on the website. The category affectation will not change during the visit, but it will be updated after the next visit.

The categorization will happen during the second visit and will last 7 months. That means that during their first visit, a visitor will not be affected to any of the Engagement Level categories. Here is the list of data we use to dispatch your visitors in our four levels:

Criteria

Explanation

visitNumber

Number of sessions performed by the user

highest amount of time on site in one session

The highest amount of time spent by the user during his last session

For example, if the user had three sessions of 4 minutes, 2 minutes, and 15 minutes each, the value of this metric will be 15.

average time on site per session

Average amount of time spent by the user during his last sessions

For example, if the user had four sessions of 4 minutes, 2 minutes, 15 minutes, and 3 minutes each, the value of this metric will be 7.

Another user that would have had three sessions of 5 minutes, 7 minutes, and 8 minutes each would have the same average of 7, but not the same highest time.

highest number of page views in a session

The highest number of pages viewed by the user during his last sessions

For example, if the user had three sessions of 4 page views, 2 page views, and 15 page views each, the value of this metric will be 15.

average page views per session

The average number of pages viewed by the user during his last sessions

For example, if the user had four sessions of 4 page views, 2 page views, 15 page views, and 3 page views each, the value of this metric will be 7.

Another user who would have had 3 sessions of 5 page views, 7 page views, and 8 page views each would have the same average of 7, but not the same highest number of page views.

ratio of bounces of the user

Percentage of sessions of the user that bounce For instance, if the user had three sessions, one session with multiple pageviews, another session with multiple pageviews, and one session which is a bounce (1-page view), the ratio of bounces will be 33%.

count of transactions made by the user

Number of transactions made during the user’s last sessions

For instance, if the user had three sessions: 2 sessions with no purchase and 1 with one purchase, the count of transactions will be 1.

This feature is useful to spot loyal visitors who tend to have a low transaction ratio because they often visit the website without buying. But in the end, they are still buying a lot more than other visitors.

ratio of transactions

Number of sessions with one or more purchases divided by the total number of sessions

average number of hits by session

The average number of hits recorded during the visitor’s last sessions

🔎 How does it work?

The Engagement Level segmentation uses a Machine Learning (ML) process to organize visitors into segments: Disengaged, Wanderer, Valuable and Loyal. The criteria to start using the Engagement Level segmentation is to have at least two buying cycles of historical data, but the more data you have available to train the AI cluster model, and the more diverse and representative that data is, the more mature and accurate the model is likely to be. The training part of the service builds the AI cluster model based on historical data from the past 190 days and updates the model every 28 days. If you have more than 190 days of historical data available, the AI cluster model should be mature enough to be accurate and will continue to improve with each update. For businesses with short buying cycles, like fast fashion, it should take less time to reach that level of maturity. However, it is also important to ensure that the data are of high quality and relevant.

The serving part of this ML puts visitors into one of the four segments, it is scheduled to run daily on visitors that have recent sessions.

It is difficult to give a specific timeline for how long it will take for an AI cluster model to mature, as this can depend on a variety of factors, including the quality of the data being used to train the model and the type of business and length of the typical buying cycle.

💡 Use case

Each category can match a specific use case:

Disengaged: you can limit the number of bounces by creating a retention modal with a special offer, targeted towards this specific segment
Wanderers: you can boost their engagement by creating a feeling of urgency and drawing their attention to the products’s low stock level
Valuable: you can activate special offers to encourage another purchase on the website
Loyal: take care of this precious cluster by offering to recognize their loyalty with a specific welcoming message

PreviousNPS® criterion (Customer Satisfaction)NextEmotionsAI criterion

Last updated 9 days ago

Was this helpful?