Unified App + Web Properties from Google Analytics

December 12, 2019
By Amanda Schroeder,
Senior Platforms Engineer, Analytics & Insight

App + Web properties have caused quite a stir — and for good reason. App + Web is not an upgrade to Universal Analytics — it is an entirely new version of Google Analytics that uses an event-driven data model designed to allow for flexibility in collecting analytics data from both web and app platforms. It also comes with an entirely new set of reports based on this new data model. 

By uniting core concepts from Universal Analytics and Google Analytics for Firebase, Google is creating a singular platform capable of working efficiently and consistently across the variety of digital platforms we use today. 

Although currently in beta, App + Web gives a clear indication of where Google Analytics is headed, and should be on everyone’s radar, regardless of what platform(s) you may be focused on right now. There is a lot of ground to cover before it really feels “on par” with the current and powerful Universal Analytics, but it’s clear there’s a lot of excitement and momentum behind this new tool.

In this post, we will cover what App + Web is all about and how it differs from Universal Analytics. We’ll also do our best to cover the most common questions and concerns that have come up around it.

Why App + Web and Why Now? 

Google Analytics has always evolved, adding new features like Event reporting, Ecommerce, and Enhanced Ecommerce. With Universal Analytics, Google attempted to take its existing product and make it work across more platforms, with mobile-specific views and SDKs, and the robust Measurement Protocol.

Despite these changes, the website-oriented nature and constraints of Universal Analytics just don’t fit the mobile app environment in many cases. These attempts at bringing web and app data together inside Google Analytics essentially were designed to retrofit website-oriented Universal Analytics for a mobile app, which in practice wasn’t always ideal (Google sunset the GA mobile app SDKs starting in October 2019). 

There are conceptual differences (e.g. a session on a website is often entirely different than a session on a mobile app) as well as technical differences (e.g web tracking expects “pageviews” while app tracking expects “screenviews”). Add smart TV apps and other platforms into the mix, and the situation becomes even more uncomfortable.

In 2016, Google launched Firebase Analytics, which was later rebranded to Google Analytics for Firebase and became the recommended approach for trackings apps. Google Analytics for Firebase’s event-driven data model is more flexible and effective for working across various platforms because it is less prescriptive and far less complex.

Now, with App + Web, Google has taken the new event-driven data model and the benefits of Google Analytics for Firebase and made this available to use on websites as well, allowing for a more flexible analytics architecture that truly works across multiple platforms. 

Below are just a few examples that dive deeper into these concepts and the benefits that App + Web properties provide.

If you’re looking for what to do next, feel free to skip to the end: App + Web: What Should I Do Now?

Holistic View of Customer Base

The ability to easily get a holistic view of users across website and app platforms has been a long-desired goal for many of us in the industry. App + Web makes this a reality, as you can configure multiple web and app “streams” into one property/dataset.

Consistent Cross-Platform Reporting

One of the primary goals of any cross-platform analytics implementation is consistent, comparable reporting. This has historically been extremely difficult to accomplish considering all of the various types of technologies in play today. Having a platform that uses a data model flexible enough for each platform is the first step, and App + Web starts to solve for this.

Flexible Reporting

Adapting the Google Analytics for Firebase data model for App + Web results in a simplified, event-driven model. This model frees us from the Google Analytics construct of everything being tied to sessions. A simplified data model means that it’s easier to share numbers internally that tie to tangible concepts. The impact of this is significant — for example:

  • Fewer scoping concerns and confusion when building custom reports
  • Segmentation by events, not just users and sessions
  • Increased emphasis on users and their events

In addition, Advanced Analysis is available for all App + Web users that enables more custom reporting and funnel analysis. This feature was previously only available to Google Analytics 360 customers. Reports in the App + Web UI are customizable, including the attribution model.

Greater Analysis Potential with Google BigQuery

A huge benefit to using App + Web will be the BigQuery integration, which is free for App + Web properties. Where previously this benefit was limited to Google Analytics 360 customers only, this benefit opens up a lot of new opportunities for all users to analyze their data in a different way, integrate their tracked data with internal data systems, and connect with other tools from Google and third parties.

The new data model greatly simplifies how the data is stored, which also means simpler SQL queries in BigQuery, shortening the learning curve for new users that are new to BigQuery. Setup isn’t the most intuitive, but you can enable the link with the Firebase console.

Simplified Implementation

Events and User properties — that’s it. All Events have the same format. All User properties have the same format. This results in far simpler implementation, and less effort required to align with the data model. It may seem strange at first, but developers and new users will be able to grasp the concepts easier. It will take an effort to put together your measurement strategy.

Key Differences from Universal Analytics

We have been working with the App + Web beta (and Google Analytics for Firebase) for some time now, and have come across some interesting opportunities and potential pain points with our customers — we will do our best to outline the most noteworthy of those below.

Universal Analytics vs App + Web Data Model

One of the key differences between App + Web and Universal Analytics is the way that data is collected, processed, and stored. 

Universal Analytics uses a data model that looks like this: 

universal analytics data model diagram

Users can have any number of sessions (visits) and within those visits, they can have any number of hits. Hits are broken down into pageviews, events, and technically, social. In this model, sessions are a primary focus of the way the data is organized and how it is presented in the reports. 

App + Web (regardless of whether you have an app or are only using it on your website) uses a data model that looks like this:

App + Web Data Model diagram

Sessions are no longer part of how analytics calculates and stores the data. That doesn’t mean that they’re gone for good — there is still a way to report on sessions. App + Web still determines how many sessions a user had based on certain thresholds of hits within the same time period, and makes this is available for reporting (the definition is very close to that of Universal Analytics with a 30-minute session timeout and customization options available). 

Events are a little different in App + Web than they are in Universal Analytics. When sending in Events in App + Web, Events are always formatted the same way: just an event name and then up to 25 custom parameters associated with them. There are no dedicated Event Category, Action, and Label fields.

Custom parameters can have any name you want. Think of them more like hit-scoped custom dimensions from Universal Analytics, though they can vary between Events rather than being defined at the property-level. This means immense flexibility to amount and type of data that can collect. 

There are no additional hit types in App + Web (screenview/pageview). There are no transaction hits or social hits. 

There is a feature similar to user-scoped custom dimensions from Universal Analytics — these are user properties. User properties have a name and a value, and you can set up to 25 of them currently at the App + Web property level. Once you set a user property, it persists for that user until its value changes or is set to null. 

There are limits to App + Web that can make implementation a challenge right now. This is especially true if you’re a Google Analytics 360 customer who is used to having 200 custom dimensions per hit with Universal Analytics. Currently, the limitations of App + Web more closely align with those of the standard/free version of Google Analytics. Keep an ear out for updates here as the product grows.

Note: There currently is no limit to the number of events logged. There is a limit of 500 unique Event Names. If you wish to capture more than 25 parameters per event, consider triggering multiple events or a creative approach to joining additional metadata with event data via BigQuery (rather than capturing it on Events). The most recent version of limitations to App + Web can be found on Google’s support site, Collection and configuration limits.

Reporting Difference between App + Web and Universal Analytics

More often than not, the limited reporting is a top concern for customers. As an early beta, App + Web still has a way to go to compare to Universal Analytics. We are looking forward to seeing growth here, and outlined several key benefits above, but for now, here are a few things to be aware of:

Rumor has it that App + Web will also offer a much faster reporting interface and additional improvements to capabilities with real-time data, especially given its simpler data model and less reliance on processing.

Measurement Protocol TBD

The Universal Analytics Measurement Protocol enables tracking of offline activity and conversions. It also is a critical component to tracking applications where there is no Firebase SDK support. 

In the background, App + Web uses “v2” of the Measurement Protocol — but it is not officially supported yet for the purposes mentioned above. We are looking forward to a solution here as it will be a key component of truly enabling cross-platform analytics for customers with various types of applications, who really need App + Web the most.

Filters/Views

Currently, there are no separate views or ability to filter data from App + Web properties. This poses a few challenges currently, such as:

  • Cannot exclude internal traffic
  • Separation of production data vs. data from development environments is tricky if you have both a website and apps (create separate streams for dev vs. prod, which means separate properties for web but you would likely have dev and prod streams for your app in the same App + Web property)
  • Dual tracking, or sending data from one application to multiple properties,  is not easy to do on the app side because Firebase doesn’t support it, but it can be done on the web side

On a positive note, App + Web by nature enables “rollup” reporting without requiring 360.

Integrations

This is a little tricky right now due to the differences between the App + Web property and Google Analytics for Firebase on the app side. App + Web is essentially Google Analytics for Firebase with the addition of web tracking (which wasn’t an option for Firebase in the past) — so some configuration options are available in the Firebase console but not yet available within the App + Web property. 

Ad Platform and Third-Party Integrations:

  • None for web/within the App + Web property
  • Potentially use the Google Analytics for Firebase integrations as a workaround (within the Firebase console), or just wait for the integrations to become available
  • Salesforce Integration is not yet available

Data Import

Data import is a feature available to Universal Analytics, which allows you to import data from external sources into Google Analytics for use in analysis. Data import is not available to App + Web, but the free integration with BigQuery allows for some flexibility here in joining analytics data with external data sources.

App + Web: What Should I Do Now?

App + Web presents an entirely new version of Google Analytics, and it is worthwhile to learn about and even start testing — regardless of whether or not you have apps. If you do have mobile apps or other types of applications and have a significant need for cross-platform analytics, we’d recommend starting to implement and test immediately. 

App + Web properties are available publicly today. However, the product is still in beta, so it is not quite ready to be your primary analytics tool. We recommend running App + Web in parallel with your existing setup for now

If you are using Google Tag Manager on your website, adding the App + Web web tracking code is pretty quick. On the app side, if you already have Google Analytics for Firebase installed, you’re good to go once it is upgraded to your App + Web property.

Stay Up-to-Date

Keep an eye out for upcoming blog posts and webinars from Bounteous with more information about App + Web. There are plenty of other resources out there as well, including the implementation guide in the Google Analytics developer portal.

Check back frequently - We’ll continue to put out resources about App + Web, including implementation instructions and product updates!

Attend our webinars - We’re kicking off 2020 with a series of webinars on the new tools, its benefits, and Q&A. Join us in January or register to receive the recording.

Join our in-person training - We’ve been teaching Google Analytics for over a decade! App + Web has been incorporated into our Google Analytics trainings, and look for specialized courses in 2020.

Look for new information and announcements over the next year from Google and industry blogs, like this one from Google in December 2019 that announced new reports and automated insights.