Session Quality measure

by Benji Azaria

2017-07-27

Google Analytics recently launched a new metric: Session Quality
This new measure calculates a score for each session reflecting the user’s proximity to conversion using similar machine learning to that of Smart Lists and Smart Goals. The new metric surfaces in a new “Session Quality” report in GA and is also available in GA custom reporting, segmentation, and remarketing.

To locate the Session Quality report go the Behavior menu that’s under the Audience report:
Locating the Session Quality reports

Here you can find two types or reports

1. Session Quality as a Dimension

Session Quality in a report as a Dimension
This is actually the only place in GA you can see this metric as a dimension ans not a metric.
We see here 4 grouped levels of quality, each group we can drill down.
Every session receives a score from 1 to 100. A value of 0 indicates that the metric is not calculated at all, probably because you’re viewing old dates.
Sessions with high value (closer to 100) are sessions with higher chances to perform a transaction.
Sessions with low value (closer to 1) are sessions with lower chances to perform a transaction.
It’s easy to understand that also with this report, you can see that sessions with low Quality Score have almost no transactions.
Sessions with high Quality Score almost half of them performed a transaction.

2. Session Quality Score as a Metric

Session Quality in a report as a Metric
Here we can see by traffic source the Session Quality. We can obviously add secondary dimension to drill even deeper.
Our best way to score a performance of every traffic source was by adding goals and viewing under the Acquisition reports traffic source and goal conversion rate.
Now we have another indication which is strong especially for predicting chances of new customers to convert based on their signals they leave us in our site.

There are 3 more ways we can use the Session Quality metric

1. Custom reports

We can create custom reports with the new metric.
Notice the Session Quality is Session scope so you will probably want to view dimensions in session or user base such as: traffic source (source, medium, campaign, content, keyword), age, country, user type, device category, landing page etc.

2. Segmentation

We can build a segment either manually, or by clicking on the small chart in the Session Quality report:
Session Quality Segmentation

For example, we can create a segment for Session Quality > 50. In the example above these are the top 5% of our users.
We can now analyze their landing page.
What is their Avg. session duration or pages per session compared to other users.
Which micro goals contribute more and which less.

We can also do the exact opposite, segment users with a Quality session score of 1 and analyze all above.

3. Remarketing Audiences

We would like to help our top 5% users that are on the threshold of converting to complete their purchase. I wrote before about tracking users with many items in their shopping cart and didn’t convert using the enhanced e-commerce events. Here we can do this relying on Google machine learning algorithm.

Creating these remarketing lists enable us to bid very high so that these users return to our site and complete the purchase.
We can also create remarketing lists with poor users (that receive low score) and exclude them from our remarketing lists.

Prerequisites

In order to calculate the metric, you need to have the following:
1. Google Analytics eCommerce implementation
2. Minimum of 1000 transaction
3. Once you reach 1000 transaction GA needs 30 days to train it’s data