Google Analytics Segmentation: Updated for Better Analysis

Google Analytics Segmentation

I’m sure everyone here knows that segmentation is probably the most important tool that we have in our analysis toolbox. And with the updated Google Analytics segments you can slice and dice your data in some really creative ways.

Let’s take a look at the updated advanced segmentation tool, discuss how it’s changed and how you can use it to analyze your website or app.

What’s new with Google Analytics Segmentation?

There are three major changes coming to segmentation:

1. True User Segmentation The big news is that advanced segments now support true user segmentation. This means that Google Analytics will actually segment people visiting a site or using an app. It also means we’ll be able to do things like cohort analysis.

I should note that there is a limit to user segments. You can only apply a user segment over a 90 day period. Remember, the segmentation tool is literally reprocessing the data in real time, which is a lot of work.

2. Sequence Filters. This is actually really big too! Now that we have true user segments we can track the behavior of users across multiple sessions. For example, you can create a segment of all the people who visited from a campaign and then bought a product some time after seeing the campaign.

3. A refreshed interface. The segmentation tool has a cleaner, more logical look. Some power users may think that it’s overly simplistic. But all the functionality is still there. The new format will make it easier to create more common segments.

Here’s a quick look.

The new Google Analytics segmentation interface.

The new Google Analytics segmentation interface.

Let’s dive in and learn a bit more about how the segments work.

Understanding User, Visit and Hit Segments

User segments will return all of the data associated with a user. This means all of the visit data that the user generated during your time frame. You can apply a condition to the user, the visits or the hits.

For example, I can create a user segment based on revenue. In this case I want to see all users who generated $500 in revenue. The metric in this condition, revenue is applied to the user. So the segment will total up revenue based on each individual user.

It could be that a user generated $500 over multiple visits while another generated $500 over one visits. Analytics does not care. As long as the revenue for the user is greater than $500.

A user based segment evaluates data for a user, not an individual session.

A user based segment evaluates data for a user, not an individual session.

Visit segments will return all of the data from all of the visits that meet your criteria. This is the same functionality that exists today.

Let’s look at the same revenue example but now from a visit perspective. If we create a visit based segment $500 revenue per visit, analytics is going to look at all of the visits that matched criteria and return that data.

Visit based segments in Google Analytics evaluate the condition based on session data, not user data.

Visit based segments in Google Analytics evaluate the condition based on session data, not user data.

Finally, there are Hit level segments. These, in some way, are also similar to what we have in Analytics today. You can apply a hit-level condition, like a pageview or an event, and Google Analytics will return all of the data from the entire visit – not just the HITS that match the condition.

Creating Simple Segments

Now that we know the difference between the types of segments let’s create some simple segments.

The left part of the navigation has a list of dimension groups. Within each group you’ll find the most common dimensions or metrics in that group. For example, in the Ecommerce group you’ll notice both metrics (like revenue) and dimensions, like Product and Product Category.

The layout for the Google Analytics segmentation builder.

The layout for the Google Analytics segmentation builder.

You’ll also notice that there is conditional logic for each metric or dimension. If you’ve used segmentation in the past then these will look familiar. And there’s a place to enter the value you want to evaluate in the segment.

For some metrics, you have the option to apply a condition to a user OR to a session. That’s represented with a small drop-down next to the appropriate metric.

Let’s go ahead and create a simple segment for User Revenue greater than $500. First I specify that I want this applied to users. Then all I need to do is choose the right conditions and enter 500 into the revenue field.

Enter a value for each of your segment filters.

Enter a value for each of your segment filters.

Notice that when add a condition I get an indicator in the summary pane on the right.

If I want to add additional conditions to this segment I need to navigate to another dimension group and add the condition. Let’s try that.

I want to add a location condition to this segment. So I click on the Demographic group and I’m going to create a condition that includes people from Massachusetts.

Notice how the interface indicates that your segment includes multiple filters?

Notice how the interface indicates that your segment includes multiple filters?

Now I have a segment of users, located in Massachusetts, that have spent more than $500 dollars with my company.

I can apply this segment to any date range less than 90 days.

One other thing to notice. Use the Options link to add this segment to any other profile that you commonly use.

You can apply your Google Analytics segments to multiple profiles when you create the segment.

You can apply your Google Analytics segments to multiple profiles when you create the segment.

Create Advanced Segments

Now that we have the basics mastered let’s move on to more complex segments. Using the Advanced section I can create segments that include multiple user conditions, multiple visit conditions, or a combination of the two.

You can apply both visit and user conditions with the Advanced settings.

You can apply both visit and user conditions with the Advanced settings.

This lets you measure the impact of a single session behavior on the long-term behavior of a user. For example, you might want to understand how reading your blog impacts the long-term value of a user.

Let’s build that segment.

First, create a user filter based on the value of the user. Then add a visit filter to match viewing blog content.

You can apply both a visit and user condition to a Google Analytics segment.

You can apply both a visit and user condition to a Google Analytics segment.

The filters are applied in two stages.

First Google Analytics selects users based on the user filter (and includes all their visits).

Then Google Analytics selects visits from the resulting set of data from the user condition.

This very simple setup can help answer the age-old question of, “does my blog content help drive more sales and revenue?”

All the content marketing fans are drooling over this segment :)

Creating Sequence Segments

Another advanced feature in the segmentation tool is the ability to create a sequence segment. A sequence can be between visits or it can be within a single visit. This is very similar to the sequence segments used in the Remarketing Feature.

A sequence filter could be used to segment user who viewed two-consecutive pages in a row. Like two steps in a shopping cart.

Here’s a segment that’s interesting. People who started a checkout process but did not purchase. It uses a page condition to represent adding something to cart. Then we evaluate if transaction were equal to zero during the session.

This Google Analytics segment will identify those that start the checkout, but do not finish.

This Google Analytics segment will identify those that start the checkout, but do not finish.

Sequence segments are really useful when you need to see how two actions, either concurrent or separate, impact behavior.

Creating Cohorts in Google Analytics

I wrote a post last year about creating cohorts in Google Analytics. It was a code-based method and, needless to say, a bit hard to implement. But we can now create basic cohorts in Google Analytics.

A cohort is basically a segment that includes some type of date condition. For example, I might want to look at all of the people that first visited my site in January. What have they done since January?

First, you can create a cohort based on a user’s first visit to the site. There’s literally a new dimension called Date of First Visit. This means that you simply specify the date range, or a single date, for the first visit.

You can create a basic Cohort in Google Analytics using the First Date dimension.

You can create a basic Cohort in Google Analytics using the First Date dimension.

NOTE: The date range for the cohort can not be more than 31 calendar days.

You can also use the date of first visit dimension when creating other segments. For example, if I am a commerce company, I might want to combine the segmentation of first visit with total revenue for the user.

I would combine the Date of First Visit condition with the E-commerce condition.

I can combine a First-visit date condition with a user revenue condition in Google Analytics.

I can combine a First-visit date condition with a user revenue condition in Google Analytics.

The resulting data would look something like this:

Here I can see all the revenue data from user who first visited the site during a specific timeframe.

Here I can see all the revenue data from user who first visited the site during a specific timeframe.

Reminder, user-based segments can only be apply over a 90 day period. So any cohorts you create using this method can only be applied for the last 90 days.

Here’s another example. If I’m a publisher, I might want to evaluate how people who visit for the first time in a given month perform over time. I can create a segment for each month (first time user cohort) and see how each performs over time.

Viewing a monthly cohort in Google Analytics.

Viewing a monthly cohort in Google Analytics.

If you’d like to look at cohorts over a timeframe longer than 90 days check out this other method for tracking cohorts with Google Analytics.

Something important to know here. We can not yet create cohorts around a conversion. There are no dimensions in Google Analytics that contain the date of conversion. We only have a dimension for First Visit Date.

Making it easier to Manage Segments

We’ve covered most of the changes, but there are a few more changes that will make life easier for segmentation users. It’s now a LOT easier to manage all your segments.

You can quickly filter all your custom segments from the pre-made segments. As well as search for segments and identify your favorite segments with a star.

If you want to edit a segment click on the little gear icon that appears when you hover over the segment.

There are a number of ways to organize and find your Google Analytics segments.

There are a number of ways to organize and find your Google Analytics segments.

Wrapping up

This was a really, really long post. But I think this is one of the most important changes ever made to Google Analytics. The ability to do user segmentation really changes the way that people with use GA. And, as we move into the world of Universal Analytics where we have more data, the ability to segment users more effectively will be important.

Be Sociable, Share!

    Like this post? Sign up to get posts delivered to your inbox.

    Comments

    1. says

      Great post Justin, and I’m looking forward to trying these out. When will they be available? I go into GA right now and still see the old-style Advanced Segments.

      Thanks,
      Tom

    2. Richard Dawson says

      Hi Justin,

      Quick question: does this new user segmentation capability require the Universal Analytics tracking code or can the standard tracking code be used?

      Thanks!

      • says

        @Richard: No, segmentation sits on top of the data in Analytics. It does not matter where the data comes from. It could be an app, website, or in the case of Universal Analytics, a kiosk. Hope that helps!

    3. says

      Thanks for this awesome post. No doubt in it that with this roll out Google Analytics team just earned more respect from the folks around – you guys are doing a ***ing awesome job. On the second note, I hope these new segments are still share-able and can be applied to any profile? Do you have any insight on what will happen to existing segments when the new interface will roll out? Thanks.

      • says

        @Mohit: Thanks Mohit. GA is still getting better and better! This will roll out over the coming weeks, thanks for your patience! And yes, you can apply the segment to multiple profiles when you create it.

    4. says

      Great post and great news for all of us! We’ve been using this in Beta for a few weeks, and yes it is a major milestone for GA and provides some awesome new analysis capabilities!

      I have to say, though, I’m not completely thrilled with the new UI. It seems to be harder to find and select the segment I’m looking for, since the segments are not in a nice compact list. I typically work on a widescreen monitor, and I have to look all the way across my monitor, back and forth several times. I’ve found myself using the search box more, which I guess is OK but it requires using both the mouse and keyboard.

      I wouldn’t sacrifice the new functionality, though. Being able to analyze visitors over time is HUGE!

    5. alvaro cruz says

      I have a question that is if I want to have the GA in spanish version will it be easy to get it? And if I want to use this methodology in a spanish speaking country the criteria and the items will be easily adapted to a specific country??
      Thank yo very much

      • says

        @Alvaro: I think you’re asking if the new segmentation will be available in Spanish, correct? It should be available in all the languages that Google Analytics is available in. I think that’s 42 different languages. I hope that helps.

    6. Ulrik says

      Great post. I would assume that Google have timestamps on all their server calls and thus would be able to provide cohort analysis for any data they wanted to. The reason they do not, is quite likely a matter of how much server power they prefer to let users handle.

    7. Katja Nielsen says

      Great article and feature! I would love some short videos showing examples of how to use these cool feature. Have a great summer!

    8. Dancho says

      Excellent post Justin!!!

      While I was reading your post i tried to replicate those segments in my GA account. However, I got bit packed when I set up the sequence segment with the checkout.
      First, I created segment how many people do checkout 55%
      Second, created another segment how many people do checkout but don’t book 43%
      Third, created third segment how many people do checkout and book 0,7%. Where the rest of 11,3% go or do?

      Also, I tried several times to create basic cohort segment only by date of first visit but I always get 0.

      Looking forward to your advice

    9. says

      Hi Justin – Great post. Thanks for coming out so quickly with an expanded explanation of the GA Blog post yesterday.

      This question is going to sound pretty basic. What is the difference between GA’s traditional metric of Unique Visitors and the new option to track Users?

      Thanks so much,

      David

      • says

        @David: Thanks for the kind words, I appreciate it. Right now there is no difference between unique visitors and users. But that will change in the future. With Universal Analytics you have the option to measure users across devices. This means deduplicated user measurement. In that scenario it makes more sense to use the term user than unique visitor as visitors really only apply to websites, not apps and other devices. It’s a very suble difference but important. Hope that clarifies things!

    10. Andy P says

      Thanks Justin, Great info as always!

      When you say the user segment can only be applied over a 90 day period, do you mean 90 days back from today, or any 90 day period? For example, if I want to compare behavior YOY, can I use a segment that show 90 days back from a point in the past (say last year)?

      Thanks!

    11. shakolati says

      Can you create multiple groups of contrasting characteristics or behaviors in one segmentation? A purpose of this would be to create one segment object and see a chart with the data from these multiple groups by interacting with that one segment object.

      • says

        @shakolati: Thanks for the comment. Yes, you can create multiple conditions within one segment. This can be a combination of user behaviors or session behaviors. Let me know if that answers your question!

    12. Antony says

      Hi, any information about the underlying technology that i used for user tracking. Would you know which cookie(s) it is using, its (their) limitations because this could potentially redefined best practices for tracking (too many options right now…). Thanks in advance for your reply.

      • says

        @Antony: Segmentation does not change any of the underlying tracking technology, just the segmentation of the data. GA will still use a first party cookie to measure users on a website. For mobile apps, GA will create a unique ID and store it in a local database on the device. Again, this is still the same mechanism used today. Hope that helps!

    Trackbacks

    1. [...] These new features are extremely important to the future of Google Analytics. With the recent roll out of Universal Analytics providing us with more data, having the ability to segment users more effectively is crucial. Watch for these changes over the next three months. If you would like to learn more about Google Analytics Segmentation and Cutroni’s blog post, click here.  [...]

    2. [...] 更新后,Analytics的使用者还可以执行同期群分析。Google表示,使用者可以分析来自特定受众(举例来说,首次访问后指定时间内的受众)的长期获益。比如你想了解所有12月首次访问自己的网站,且节假日期间在上面花费超过100美元的客户的分析数据,之前这很难操作,但更新后,只需点几次,报告就会出现在眼前。 [...]

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>