With 3.8 million apps available for Android users and 2 million apps in Apple's App Store, it can be tough for an app developer to stand out among the competition. But with Google's Universal App Campaigns (UAC), developers have an opportunity to market their mobile apps with targeting options based on audience demographics and behavior. It all happens automatically -- as long as you set up the campaigns correctly. In this post I take a look at how you can put machine learning to work for you, using the power of Google’s Universal App Campaigns. [tip] You can now see cross-device activity in Google Analytics 4 as Universal Analytics Sunsets in 2023 [/tip] Campaign set up Getting started with a UAC is relatively easy. The three steps are to identify an audience, ensure conversion tracking is set up correctly, and relevant text, video, and images are available for the campaign. The two major actions for UACs are to find new users who will install the app or those who will perform an action inside the app, such as making an additional purchase. One the UAC is set-up, it is eligible to show on Search, Display, YouTube and the Play Store. The initial setup is straightforward. The advertiser only needs to provide four lines of text with images and with machine learning, Google decides which combination to show to a particular user. Goals When you consider goals for your UAC, the install action is an obvious one regardless of the app category. Targeting options includes people who are likely to install the app or who are likely to install it and perform in app action. It is up to the advertiser to determine what a valuable action looks like and ensure conversion tracking is set up before launching a campaign. In-app actions, or goals, or can be either success actions or proxy actions. With a success action, the app user makes a purchase inside the app, upgrades the service, or signs up for a paid subscription; something that generates revenue. Assuming success actions happen at least ten times a day with users, the system has enough data to identify and target the right audience for your UAC. If volume of success actions is low, there is not enough data for machine learning to make decisions. In that case, the advertiser can identify a proxy action which is a behavior that is likely to lead to success action. An example of this is someone who added payment information to upgrade service but did not follow through with upgrading. Or it could be tracking which of your users share incentives with their network. Advertisers need to think carefully about what a proxy action truly is. When it it is too early in the funnel, it includes people who are less likely to convert and not a good representation of those who will later perform a success action. If a mid funnel behavior is identified as a proxy action, rather the the top of the funnel, it may better represent people who are closer to converting so it is more likely to later result in a success action. [subscribe] Conversions Setting up and collecting conversion data is a crucial piece to success because these campaigns look at past searches, browsing behavior, and other apps used to determine who is most likely to convert. Before launching a UAC, ensure this conversion tracking is set up correctly or your will not be measuring goals that matter. For e-commerce sites, the primary conversion is clearly to drive revenue in the form of an in-app purchase or perhaps subscriptions. With luxury retail, it is especially important to have conversion recording correctly because of the multiple touch points. And Shopify users can use the Littledata reporting app to gain even more insight on the user journey through that platform. Measurement and optimization There are immediate metrics to monitor - app installs and in-app purchase - but there are also long term considerations such as the customer lifetime value (CLV), that should be part of your overall strategic marketing plan. A single user who makes a purchase provides direct revenue. If they refer someone to your app, that is considered indirect revenue. The first number is clear-cut revenue and easy to measure. The second is one that you determine based on your internal data, meaning what type of behavior and interaction with customers generally leads to a sale. The value of both of these actions contribute to the CLV. Lifetime is the length of time they interact with your app. If they install the app and use it to buy things over the course of a year, then stop, their CLV time period is one year. Once you have identified your CLV, use this value to set your target CPA and optimize it based on performance. Decide what you are willing to pay for a success action and what you will pay for a proxy action, knowing that number will likely change over time. As data comes in from your UAC, you can compare the lifetime value of your different customers through segments. Segments help you uncover those customers who purchase every couple months compared to those who only make an initial purchase. Those the make multiple purchases represent segments with a higher value. Drilling into data with segments allows you to see who gives you the best return for your investment. This level of detail helps you identify how much you paid in your UAC for to acquire each type of customer so you can adjust accordingly. Review what you paid initially for the type of users that you bring in and compare that to their lifetime value. Are you investing your budget in a UAC that brings in users that generate recurring revenue? When you bid strategically based on a lifetime value, you are not overly focused on short-term transactions. It is less expensive to keep a customer than to acquire a new one so you want to think in those terms. What next? Decide on UAC goals that make sense for the purpose of your app. What should users do in addition to downloading the app and what behaviors indicate they are getting close to a conversion? Gather assets - text, video, and image - that are enticing for users and ensure conversion tracking is setup properly. Without proper conversion tracking, you miss out on the data you need to determine success. Monitor performance of your campaigns, and if you run an ecommerce site, track a wealth of data with the Littledata app. Think about the CLV and optimize your campaigns to reach the right users rather than any users. Your bottom line is generating revenue so keep that in mind with every UAC. With careful planning and well managed campaigns, your app can stand out in a crowded marketplace.

This is the third article in the Q&A series. I will be answering some of the most-asked questions about Google Analytics and how it works. If you’ve missed the previous articles, you can access Part 2 (What is the bounce rate in Google Analytics) and see what questions we answered there.   In this article, I will give you an answer to the following questions: How Google Analytics works for mobile apps? What are the differences between Firebase Analytics and Google Analytics? How Google Analytics works for mobile apps? Instead of using JavaScript, for mobile apps, you will be using an SDK. That is a Software Development Kit and it’s what collects the data from your mobile application. As most smartphones are either Android and iOS based, you will have different SDK’s based on the operating system. The SDK works similarly as the JavaScript and collects data like the number of users and sessions, the session duration, the operating system, the device model and the location. All of that is packed in hits and sent to your Google Analytics account. Here is an overview from The Google Analytics Help Center. The main difference is that the data is not sent right away. Because a mobile device might not have a connection to the internet at some points in time, the data is stored on the device and is sent when it is eventually connected. The process is called dispatching and it’s done at different time intervals on Android and on iOS. On Android, the hits are dispatched every 30 minutes and on iOS, every 2 minutes. Those numbers can be customised though. [subscribe heading="Need help with Google Analytics?" button_link="https://www.littledata.io/contact-us" button_text="SCHEDULE A DEMO"] Keep in mind that you can customise the code so that you can track different data in case you feel the base code is not sufficient for you. What are the differences between Firebase Analytics and Google Analytics? Firebase Analytics (FA) is another way to collect the event data. While Google Analytics is a general-purpose (and more web oriented) analytics tool, Firebase was built keeping mobile in mind. There are some things that were added in in the later and also things that are missing from GA. Here are some noteworthy points when considering Firebase Analytics: Real-time view is missing for Firebase Analytics (we mainly use this when testing the app for new events). Events are available after 4 to 6 hours in Firebase Analytics. The Behavior Flow is missing from Firebase Analytics (since there are no screen views logged). The Audiences feature is a big advantage that FA has. If you couple this with the Notifications it will allow you to engage with a specific group of users. If users experience a crash, then an audience group will be created automatically when using the Firebase Crash Reporting feature. Funnel analysis based on custom events is easier in FA. However, if you use Littledata, then this problem can be solved for Google Analytics with the custom reports that we can build. Some events are logged automatically in Firebase Analytics (for example the sessions based on the Activity life-cycle). Firebase has a relatively low methods footprint compared to the methods count that Google Analytics uses - making it less processor and network intensive. As a final point there are benefits for using both platforms to track your Analytics, but if you do focus your business on mobile applications, keep in mind that Firebase Analytics was created for mobile apps. Happy Reporting. Get Social! Follow us on LinkedIn, Twitter, and Facebook and keep up-to-date with our Google Analytics insights.