Mastering Sequential Logic in AA & CJA: Introduction to THEN

[#1]

Do you know how sequential segmentation works in Adobe Analytics (AA) and in Customer Journey Analytics (CJA)? Not something basic (left) but something more interesting (right). Do you know what happens in this segment using Only After Sequence, plus Exclude, with multiple checkpoints, out-of-order container levels, and an After modifier? Do you know why this is valuable?

That sample belongs to a use case to explore a visitor’s next visit after some key event, like an application submission. It returns exactly one visit for everyone who has applied - their next, consecutive visit. It works for the whole dataset, though qualified visitors applied on different days and engaged again at completely different times!

There is so much I want to tell you and unlock for you regarding sequential segmentation in AA and CJA! This is the first post of my series Mastering Sequential Logic in Adobe Analytics and Customer Journey Analytics. I’ve specialized in this subject for many years at Adobe. My aspiration is to enable you to use these features confidently and to ask questions of your analytics data that you thought were impossible! We will cover the example above and many others, starting with foundations and then building to advanced scenarios.

Series table of contents

Level	Topic & link	Post date (actual or estimate)
Foundations	Introduction to THEN (this post)	Feb 19, 2025
Foundations	AFTER / WITHIN	~ Mar 7
Foundations	EXCLUDE	~ Mar 28
Foundations	ONLY BEFORE/AFTER SEQUENCE	~ Apr 18
Intermediate	Visual framework...	TBD
Advanced	Advanced use cases...	TBD

Some material has also been delivered as live webinars presented by Adobe Customer Success:

• Recording of Mastering Sequential Logic in AA & CJA: Foundations (Oct 16, 2024)
• Recording of Mastering Sequential Logic in AA & CJA: A Visual Framework (Jan 22, 2025)

Foundations

Let’s start with the basics, how each core element of sequential segmentation functions in isolation. I will provide sample use cases and screenshots wherever possible to keep things real. As the series continues, I can build on the previous material and share more powerful themes and use cases. As part of this foundational beginning, we’ll cover:

• THEN
• AFTER / WITHIN
• EXCLUDE
• ONLY BEFORE/AFTER SEQUENCE

The logic for segmentation evaluates equivalently in both AA and CJA but the default labels are different. Since I’m addressing both solutions, my samples and language will show a mix of each. (In CJA you may relabel these containers to suit your environment).

Lastly, whichever tool you use, the capacity to study behaviors across touchpoints depends on matching visitor or person identifiers across data sources. There are some ways to approach this in AA and many more such capabilities in AEP and CJA. Every example assumes that we are exploring a unified dataset, whether analyzing one large web and app report suite or a unified cross-channel connection in CJA.

Enough! Let’s go!

THEN

Conditions or checkpoints met in a specified order.

Sessions where both Web and Email occurred, where Web precedes Email.

All event data in AA or CJA is grouped by visitor identifier and ordered by timestamp. So, at a minimum, the conditions separated by Then must occur on separate event records or hits. The Then operator is not even available until at least two conditions exist in your definition.

Sample business questions

You can address business questions like these with simple Then logic. This isn’t very exciting yet, but we will reach cooler scenarios later!

• Which customers purchased after using the mobile app?
• What else happens during visits where a conversion occurs prior to a marketing touch?
• Let’s study applicants who open an account before calling a branch. Then we can compare them to the segment that opens an account following call activity.

TRUE at least once

If at least one sequence matches in your session or person data, the logical result is TRUE for the whole session or person container. This is just like a segment of web visits with Page = Home, requiring only one page to be Home to then match and retrieve the entire visit.

So, our example Sessions with a Web event THEN an Email event will match sessions like:

The sequence definition must only be met at least once somewhere within the container scope.

Interface

Let’s look at the available controls within a sequential segment or filter. We will review each of these options in detail, their impacts alone and in combination, throughout the series.

Examples

To wrap up this introductory post, let’s explore two examples from within Workspace. When validating segment definitions, examples are crucial to success! Here, I’m using dummy data, but the approach is the same with actual data:

1. Find or create some activity that should or should not match your logic.
2. Focus in on that example only (often with a panel segment).
3. Compare results with and without your sequential segment applied.

Person A

Here, I’ve identified some sample data with panel filters, looking at Person A specifically. I picked this person because they only have one session of activity. It is easy to view a small example and observe whether our segment logic captures or drops their single session!

On the left, I’ve used some Workspace tricks to help visualize their activities. The rows display their events in sequence, while the columns show whether their action was on web, app, or email. Color highlights emphasize these sequential stepping stones of their session.

On the right is the same table but with our filter applied: Sessions: Web > Email. It matches this person’s data because we see all their actions in the session across web, app, email are returned. Said another way, there is no change in result, so the segment must have matched all of it.

Note that the logic evaluation can consider multiple sequences that match. For example, here there were four variations of a Web event followed by an Email event. Only one matching sequence is needed, however.

Person D

Here is another sample visitor Person D applied to the same left and right views as before. In this case, see how our result table looks when the segment logic does not match their visit. It’s easy to see why; in this simple case, the only email activities (events 1,2) occur before any web engagement (events 5,7,9). As there is no visit or session with Web followed by Email, the visit does not match our criteria, and we get nothing returned on the right.

Signoff

There is more to come (a lot more) and it gets more exciting, quickly! It’s exciting for developing business insights and subsequent value, plus because its fun to have wizard-like control over your data and analysis. The basics of THEN now behind us, next we will confirm the functions of AFTER/WITHIN and EXCLUDE (in a sequential context).

The series table of contents above will stay updated as new posts go live. It will be a guide to the entire sequential segmentation topic, by the time we’re done. See you again soon!