Google Analytics is introducing a number of new technologies, collectively called Universal Analytics, that will create a better way for businesses to measure their digital world.
The cornerstone change in Universal Analytics is that Google Analytics is becoming user or customer centric rather than visit centric. This paradigm, sometimes called customer centricity, more closely aligns with how businesses traditionally measure their performance.
Let’s take a look at the new technologies and how they can be used.
The Measurement Protocol
The Measurement Protocol is a standard way to send data to Google Analytics. The new Measurement Protocol defines how you can send to data to Google Anlaytics from any system or any device.
Data is still sent via the __utm.gif
image request and you can send data via a GET
or a POST
. As long as long as you format the data according to the protocol Google Analytics will accept the data.
This means you can send data from call center systems, point of sale systems, etc.
See the power in this?
There are a number of code libraries, in common languages like PHP, Java, etc., that let you integrate Google Analytics into other systems. Or you can build your own library. You decide what the data you want to send and then use the libraries to send the data.
There will also be a new library for Flash. The current Android and iOS SDKs already suppor the Measurement Protocol.
Speaking of libraries, the Measurement Protocol includes a new JavaScript library named analytics.js
.
Yes, you will need to retag your site. Luckily there a new, free tag management tool that helps you manage your tags more easily.
NOTE: analytics.js
supports all of the standard Google Analytics functionality, like events and ecommerce tracking. But it does not support content experiments or the new remarketing feature.
Server Side Sessionization
Sessionization is the process that any analytics tool uses to identify users and group the sessions they create. In the old Google Analytics sessionization was handled in the tracking cookies. The cookies store things like session length and campaign information, as well as an anonymous identifier.
With Universal Analytics all of the sessionization happens at the server level. The new analytics.js
does not maintain any tracking information (other than an anonymous identifier).
There are a couple of really cool benefits here.
First, almost all configuration settings that are normally done using JavaScript will be moved to the admin section of Google Analytics. This includes things like:
- Adding new organic search engines
- Setting the tracking cookie timeout length
- Reclassify certain keywords as (direct)
Another cool part of the measurement protocol and server side sessionization is the ability to use your own anonymous sessionization key. This data is an anonymous ID that Google Analytics can use to aggregate all of the data.
For example, if you have an anonymous CRM id then you’ll be able to use this in analytics and GA will group all the data using this ID.
Offline Conversion Import
Using the Measurement Protocol you’ll now be able to send offline conversions to Google Analytics. For example, if a user converts via phone you’ll be able to track that conversion in Google Analytics.
Here’s how you would implement offline conversion import:
You need to send a hit to Google Analytics when the offline conversion happens using the Measurement Protocol. When you send the hit you’ll need to include some identifier that links the offline conversion to the data in Google Analytics.
Offline conversion import could also be used with a point-of-sale type of system and loyalty cards. You could use the loyalty card ID as the sessionization key and the measurement protocol to collect all the data.
This is one of the Universal Analytics features that truly makes Google Analytics a more business-focused tool. Heck, it’s universal.
By having all conversions in Google Analytics we’ll have a much more accurate view of revenue/conversions and the overall performance of the business.
Is it possible that this is the big data moment for Google Analytics? That’s a discussion for later :)
Custom Dimensions and Custom Metrics
As the name implies Custom Dimensions and Custom Metrics are custom data that you can add to Google Analytics. To put it simply, this is the evolution of Custom Variables. You still need to understand things like the variable scope and index! So make sure you read up on mastering custom variables. Until there is a mastering custom dimensions post :)
The big difference is that Custom dimensions & metrics are set up both in the code and in the Google Analytics admin section. This is possible because of the server side sessionization.
First, you need to add code to your site to create the custom variable. Something like this:
ga("set", "dimension1", "Custom Dimension 1");
Then you need to set up the custom dimension or metric in the admin settings. That’s where you’ll configure things like the scope and manage which custom dimensions and metrics are active.
All of this server side configuration is made possible by the server side sessionization. Less coding – it’s a good thing.
Custom dimensions and metrics will be available in your Custom Reports and Advanced Segments. Custom dimensions will also be available as secondary dimensions in your standard GA reports.
I’ll have a full post on using custom metrics and dimensions once they start rolling out to more people.
Dimension Widening
Dimension widening is based on custom dimensions and metrics. The key difference is that this is a bulk-type of feature. Whereas custom dimensions and metrics are done one at a time, dimension widening allows you to add a bunch of new dimensions via a data upload.
This is a big deal for enterprises and large analytics implementations.
Let’s say you are a publisher, and your analytics data has thousands ISBN numbers (basically unique book IDs) in the content reports. You can import detailed information like publication date, author, edition, etc. into Google Analytics, link it with theISBN number and view other info like publication date, author, edition, etc.
Let’s say you’re an ecommerce site. You could import data from a product database and link it with information in Google Analytics.
Some of you old timers will remember this feature from Urchin. It was called custom lookup tables. This is a bit more advanced, but the same general idea.
Cost Data Import
A great way to use the custom metrics feature is to import cost data associated with a marketing campaign. Using an ID number in your content data (like a query string parameter named CID), import cost information about your marketing campaigns.
This is a feature that has been long-missing from Google Analytics.
We’ve always had cost data about Google AdWords but not other marketing activities. This has made it hard to measure true marketing ROI within Google Analytics.
Using cost data import we can get a more accurate view of campaign performance and the ROI of marketing activities because we’ll know how much a campaign costs along with how much revenue it generates. And that’s a good thing.
What Does Universal Analytics Look Like?
A lot of this post has talked about the different technologies and what they do. But what are some of the tactical things that you’ll need to change? And what does Universal Analytics actually look like?
This is a massive change to the platform. It enables a lot of new data. But the front end does not change drastically.
You’ll see new dimensions and metrics if you use Custom Dimensions or Custom Metrics. And you’ll see some small changes to the admin settings due to the server side sessionization. But that’s really it.
Applying to the Universal Analytics Beta
This is a beta launch – so it’s rolling out slowly. Use this form to sign up for the Universal Analytics beta.
If you are white-listed for the beta you’ll need to create a new web property to use Universal Analytics Remember the data processing is fundamentally different. It’s visitor based, not visit based.
Next you’ll need to send the new web property ID to Google and it will be white-listed.
If you’re just tracking a website (mobile or desktop) then you’ll need to update from ga.js
to analytics.js
. That’s the only change.
But if you want to integrate offline data, like a call center, you’ll need to integrate the Measurement Protocol.
Wrapping Up
There’s a LOT to Universal Analytics. Moving to a user centric system means that businesses can get a better understanding of their marketing investment and performance of content across the digital world.
I know this post may be a little vague, but this is going to be a gradual migration process. I’ll post more as time goes on and more people have access to Universal Analytics.
Just remember that Google Analytics is a platform that you can build on. It will help you execute a measurement strategy that is more closely aligned with how you do business.
Ho hum. What are you announcing Tuesday? ;-)
Cool stuff. Can’t wait to see all the ways orgs will start to use these new capabilities.
Much much awaited updates and a great move by Google Analytics team. Bravo everyone!
Thank you for this very interesting article. I’m really looking forward to testing it.
Do you know if everything will be free or will it be merged with a Premium access?
@Sebastien: Universal Analytics will be available to all users.
Its a pleasure to see, that Tools for heavy analysis of Google is followoing the development of the progressing internet world. In my opinion, it is important to have a huge variety of tools to analyze Goggle´s affairs to undertsand Google behavior and the internts development, compared with an easu access for the competence level, at which you do internet marketing or research or maybe programming.
There are still a need for excellent supporting tools for doing online work, seen from different metrics.
Lasse
It’s great to hear that Google Analytics is taking a step towards bringing more necessary data into the reports.
But there’s still some way to cover in bringing real-world events into your metrics. I mean, offline conversion and custom dimensions and metrics are still internal to your business, and are “traditional” in the sense that they’re a part of your business logic flow.
The problem starts when events happen *outside* the scope of your flow, but still have a significant impact on your business, which GA cannot currently track or visualize in your reports. To give a current example, how does Hurricane Sandy affect your conversions?
Disclosure: I am the founder of Chartelligence.
@Oren: Great point! I don’t know of any analysts that don’t like to use external data for more context. I can only imagine the productivity of San Francisco suffered yesterday after the city was up all night celebrating the Giants’ win.
One think I like about Google Analytics is the engaged community, especially from our developers. They’ve help build more than 50 Google Analytics Applications. Some of these make it easy to integrate Google Analytics data with other data sources.
Thanks for the top on Cahartelligence, I’ll check it out.
Thanks! You’re right about the community and the tools it provides, it helps Google Analytics take steps outside the box, and really be the trend setter in the Analytics market.
I also just submitted Chartelligence to the Google Analytics App Gallery, I think the community can benefit from it.
Justin thanks for this rundown. All of these things sound awesome. Visitors! I think I might just wander the halls of the Howard Space Center saying “Visitors!” for a few days.
Quick question: analytics.js will be replacing ga.js for Universal users or can it run in addition? The reason I ask is, of course, Experiments. Or are we expected to use different tools to run on-site optimization?
Thanks again!
@Gahlord: Eventually everything will run through analytics.js. For now you can run both systems together. You’ll obviously need to double tag, but that should keep the data consistent in both traditional GA and Universal Analytics.
Great post – this is really a game changer. I think it will be hard to get buy-in from ‘traditional’ businesses in the early stages. Someone will have to do it first and share the data on their success before the ‘traditional’ businesses will follow. For businesses that make data driven decisions, this is a no-brainer.
Great news! Loving the attribution modle customization and data importing!
@Jeff: Glad you like it Jeff. I should mention that all of the Attribution Modeling stuff is not part of Universal Analytics. It will be rolling out soon.
Will these new conversion touch points be available in the multi-channel funnel reports? I’m not sure if you mentioned that already. It would be really cool to see if a visitor clicked on a PPC ad, came to the site organically a week later, came to the site directly the following day, signed up for an email, came back to the site from an email link, called a sales representative and converted during that call. This would provide the visitor’s entire conversion story.
Looking forward to learning more about these new features and testing them. Great post!
@Brandon: Universal Analytics does support a lot of the ideas that you’re talking about. I’m not sure you would call it a multi-channel funnel report, that’s looking at traffic channels. We need to incorporate the device and content to that visualization as well. But yes, Universal should support a lot of what you’re describing.
Will Google (or any third-party) be offering any training programs to help analysts / developers learn about these new features?
@Andrew: There are currently no plans to launch any formal training for Universal Analytics. We usually do some webinars to explain the features but do not set up offline training. You might want to check with our Certified Partners, they usually have training courses available.
Thanks!
Hi Justin,
I think the GA community has a million questions around this change–would there be any possibility of scheduling some gchats around this after the webinars take place? Webinars are cool, but it’s usually the last 10 minutes left for questions. It’d be great to maybe have you and Nick discuss take questions (perhaps filtered through google moderator), and perhaps have a few GACPs on as well so we can understand ways they would like to/plan on using it for their clients. That last part might put you in a tough spot politically (e.g. regarding favoritism), but whatever, just throwing that out there :-)
@Adrian: Thanks for the suggestion Adrian. Don’t worry, I know there are a lot of questions and not a lot of information out there. This is going to be a migration process that takes a little bit of time. It’s not happening next week. But I feel your pain. The last thing you want to do is invest a lot of time in Google Analytics now, just to have it undone by Universal Analytics.
With that said, I’m working on a post that will detail how the system can be used, so you have some context for the features.
Hi Justin,
Any updates if we will be able to see (not provided) keywords or use some work-around to get it?
Thanks!
@Chande: You will not be able to see keyword data for users that are logged into Google. Ever. In fact, it’s not just logged in users. FireFox has changed and is now using secure search to mask keyword data from browsers (and as a result, from all analytics tools).
As a workaround, I’ve got a custom report that you can use to get some insights into (not provided). Give that a shot.
With this new version, will Google Analytics be able to track content viewed in feed readers (since the tracker runs server-side)?
@gxg: The new measurement protocol will let you integrate analytics into a lot of different systems. Distributed content could be tracked. You could, in theory, create an invisible image for use in each of your posts. You would need to create some code that dynamically creates the GA request and then embeds it in your content. Does that make sense?
Justin, I’m from Brazil and I appreciate your posts and books! Like much of his work, My question is can I put the two codes (and analytcis ga) on the same property – id?
@Christian: Yes, you can use both ga.js and analytics.js on the same page. The data is sent to different web properties and the tracking code uses different cookies. So there should be no conflicts in the data.
This is brilliant. What I’m really looking forward to is being able to bring outside data from call centers and from CRM/Marketing automation systems into Analytics rather than what we do right now i.e: mashing up data by exporting from analytics into Salesforce or Marketo. I’m playing around with the data import features right now and am very pleased with the direction GA is moving towards.
Curtoni, Is that mean universal campaign is going to come in Adwords?
Universal Analytics is the next version of Google Analytics. So all of the data will be in Google Analytics – which includes AdWords campaigns and metrics.
Thanks for keeping this blog updated about analytics, I am an avid follower of your blog, I have also implemented this integration into our CRM http://cutroni.wpengine.com/2009/03/18/updated-integrating-google-analytics-with-a-crm/ , just curious to know if you will soon publish a new blog post with Universal Analytics. Also it would be nice if you can post samples of how the cookie data can be accessed from via PHP.