Session engagement scores

EngineRoom’s Session engagement model assigns a score from 1 to 10 to each incoming lead for your business. This model is extensively trained on your customer's historical leads using various User Engagement metrics, including but not limited to:

  • Number of Pages Viewed

  • Session Duration

  • Number of Clicks

  • Time of Day

  • Session Attribution (Source/Medium)

The result is a personalised scoring system that accurately reflects the unique engagement patterns of your leads. It is only available to customer's that do not have active sales cycle data.

Let's break this down with an example

Our customer receives a significant number of leads every month but lacks the sales data needed to identify which leads convert into sales. This absence of data prevents us from training a machine learning model to predict actual sales conversions.

Instead of relying on a generic predictive model trained on other customers' data, we will build a personalised approach using the customer's own historical leads and user behaviour. Let's analyse the typical patterns observed during website sessions that result in lead conversions and develop a proxy Session Engagement Score based on these insights.

The table below illustrates the relationship between Session Duration and Conversion Rates on the customer's website. This data is essential for understanding how user engagement correlates with the likelihood of conversion.

Session Duration
Number of Sessions
Number of Conversions
Conversion Rate

0 - 30 seconds

60,000

1,040

1.7%

31 - 60 seconds

12,000

650

5.4%

61 - 240 seconds

18,000

4,200

23.3%

240 + seconds

4,800

2,,320

48.3%

Based on the above Conversion Rate distribution, we can see that sessions that spend longer on site tend to have much higher conversion rates. Therefore, any incoming lead with a longer session duration will receive a higher score in the Session Engagement model.

To sum it up...

Multiple variables are combined to build our Session Engagement Model. Each variable's input and "weight" is personalised according to your customer's historical data, ensuring the model is tailored to your specific customers. This approach allows different customers with varying behaviour patterns to effectively utilise the model for an accurate measure of session quality.

Last updated