Time-based session start analytics

session_start is a system event Bloomreach tracks whenever a device connects to a website with Bloomreach tracking enabled. Time-based session_start analytics build on the standard new vs. returning breakdown by splitting returning visitors into segments based on how many days have passed since their last session_start—giving you a clearer picture of retention patterns, re-engagement opportunities, and the overall health of your customer acquisition.

Why use this analysis

This analysis gives you a more complete picture of customer behavior than the standard new vs. returning metric. Specifically, it helps you:

• Understand retention patterns: See how quickly customers come back and how that changes over time.
• Identify acquisition trends: A spike in first session_start events points to acquisition activity; growth in short-return segments signals strong retention.
• Take action: Use the segments directly in email campaigns, UX interventions, or retargeting audiences.

What the output looks like

This analysis measures the time between session_start events in days and groups each session into 1 of 7 segments—from first-ever visits to returns after 30 or more days.

Example insight: In a mature ecommerce business, you'd typically expect a higher proportion of returns in the 1–7 day range. A high share of 1st session_start events often indicates heavy paid acquisition activity.

Requirements

Step 1: Create the expression

The expression calculates the number of days since the previous session_start. You'll use it in the event segmentation in Step 2.

1. Create the expression

1. Go to Data manager and create a new expression for the session_start event.
2. Add the timestamp attribute. This returns the timestamp of each session_start event currently being evaluated.
3. Add an arithmetic operator and then add the running aggregate last session_start timestamp. This returns the timestamp of the previous session_start event, so you can subtract the 2 to get the time difference. If no previous event exists, the expression returns no value—this identifies first session_start events.
4. Divide the entire formula by 86400 and add brackets. Timestamps are measured in seconds, so dividing by 86400 (the number of seconds in a day) converts the result to days.
5. Click Save.

Step 2: Create the event segmentation

Create a new event segmentation in the Data manager with seven segments—one for each return-time group.

📘

Tip

Copy each segment and edit its properties rather than building each one from scratch.

Click Save when all seven segments are complete. The segmentation is now available to use in analyses and dashboards.

Step 3: Create a daily traffic report

Use a Report to see how each segment contributes to your daily traffic. This view lets you spot shifts in acquisition or retention at a glance.

Example insight: If you see a traffic spike on a particular day, check which segments drove it. A spike in 1st session_start events points to acquisition activity. Increases across 14–30 day and 30+ day return segments suggest a re-engagement campaign is working.

Save the report to a dashboard to monitor traffic patterns over time.