Quick Google Analytics Trick: Use Sequence Segments to Analyse Behavior Over Multiple Sessions
September 7, 2017 -
Sequence segments are one of my favorite things about Google Analytics (available in Standard and 360). One application of sequence segments that I find not a lot of people make use of is to create sequences that cross over multiple sessions to learn about how a user’s activity in one session resulted in a specific activity from that user in subsequent sessions.
In this post, I’m going to show you a way to use sequence segments to isolate users who have performed specific actions across multiple sessions. This allows you to test hypotheses about the long-term impact that your page content is having on visitor behavior.
One thing to note is that there are a lot of attribution reports in Google Analytics that offer similar functionality, but attribution segmentation is limited to a very narrow scope of inbound traffic-related dimensions. It does not help you if you are trying to look at cross-session impact of a given on-site behavior on other on-site behaviors. For example, you may want to know whether or not a user who fills out a lead form in one session comes back in a subsequent session to make a purchase on the site. This is why sequence segments are so important.
Example Hypothesis: I think that highly motivated traffic leaves to look for a coupon and comes back from Affiliate links. If I’m right, I will negotiate with my affiliate vendors for changes to the terms of our payment structure.
How many users add an item to their cart, leave without buying, then come back again from an affiliate link and purchase in that session? Knowing this will allow you to see to what extent you are paying for traffic through an affiliate program that you may have already paid for from an alternate source. You can see conversion paths in standard attribution reports, but this will isolate for highly motivated users who abandoned relatively deep in the shopping funnel.
Enforcing the second session
The second step in this sequence is important. By including attribution information in the second step, I’ve ‘forced’ the segment to only capture the journey of users who added to cart and didn’t purchase and then came back in another session, in this case attributed to the affiliate program. This gives us both the highly motivated behavior of someone who added to cart in the first stage (but who did not buy) and their return via affiliate.
Taking this a step further
That’s all very nice but this only really gives you a sense of how many times this pattern happened. The real meat here is knowing where the original highly-motivated traffic was coming from before affiliate took credit for the sale and what the impact of that pattern was on your revenue.
1. Where did the traffic come from originally?
To find out where the original traffic source was from, you need to maintain your sequence and layer on a condition. This condition will isolate the segment to the session where the add to cart/abandon happened. To do this, open the ‘Conditions’ tab above ‘Sequences’. Take the conditions from step 1, scoped at Sessions. This will restrict to only the sessions where the first sequence obtains.
Original Traffic Attribution
2. What impact did this have on ecommerce revenue?
Similarly, if you build another two-container segment that isolates for the session in stage 2 of your sequence, you’ll be able to see how much revenue is associated with the outcome of this behavior. If you just used the first segment you’d see all revenue associated with users of this type in your time period which is likely too broad a scope for answering the specific question at hand.
Associated Ecommerce Revenue
And that’s how to use segments for cross-session analysis.
h/t to Charles Farina for his excellent post on hit-scoped segments for inspiring this one.