What is Mobile ad attribution? An introduction to app tracking
Product Content Strategist
Feb 4, 2019
Attribution is how marketers understand the journey you take to arrive in their app and what you do once you’ve landed there. When done right, there’s a data point for each of the actions a user takes on the journey, from clicking an ad to making a purchase.
But there are real challenges: a lack of industry-wide standards (rules for dropping data points), competing mobile app attribution models (disagreements about which data points count), and user journeys that touch multiple platforms. And there’s the issue of widespread fraud.
This is our first post in our series asking an important question: how does mobile attribution work? Today we’ll answer this from a technical standpoint, diving into how we find those data points and the information they contain. Whether you’re a mobile marketing beginner or simply wondering how cookies work for mobile, welcome! You're in the right place.
How does mobile attribution work?
So why is it important to run with an attribution provider and not just rely on something like Google Analytics? The most important reason is that implementing a mobile app tracking SDK enables you to make well-informed business decisions in real time. An attribution provider gives you a platform to discover where your users come from - if they arrived in your app via a video ad, for instance. We're then able to help you understand how that user moves through your app and how you can compare their journey to someone else who arrived via a different source.
This lets you determine which are your best-performing campaigns, so you can pinpoint the most effective ads and iterate on them. With this information, you’re able to optimize your creative assets and use hard data to get rid of failing ads and tweak the good ones. Greater knowledge about how your ads perform allows you to practice smart retargeting and build campaigns targeted. For example, you could specifically target users who tried out your app but didn’t stick around.
Your users will come from multiple advertising channels. If you cannot track the how, who, when and why of their journey to your app, you cannot know which of your networks are delivering users, the relative value of those users, or how much of your marketing budget is going directly towards fake clicks and fake installs.
Working with a mobile app attribution provider like Adjust gives you an eagle eye view over your data - a single, streamlined dashboard you can utilize for in-depth analysis without having to collate or wrangle data from any of your partners. Adjust also gives you the power of verified purchases and fraud protection, to make sure your data stays clean and reliable.
What happens when I click on an ad?
Let’s say that you’re using your iPhone to play a game. A video ad pops up within the game. You watch the video and click the call to action (CTA) to download the app at the end of it. The link takes you to the app in the iTunes store, but briefly redirects you through Adjust. This takes a fraction of a second but is a key step; it’s how the attribution provider receives the first data point - the engagement with the ad.
By clicking the link, going to the app store, downloading the app and opening it for the first time, the attribution provider will receive the following data points:
- Advertising ID - a string of numbers and letters that identifies every individual smartphone or tablet in the world
- IP address – a specific address that devices use to communicate with one another via the internet
- User agent – a line of text that identifies a user’s browser and operating system
- Timestamp – When you clicked on the link
- First Install - When you first installed and opened the app
With this information, the attribution provider can determine whether the user is new or existing. If the user is new, the attribution provider will attempt to match the user’s install to their engagement with a particular ad. This exchange of information can happen in several ways; the most common is for the app to integrate the attribution provider’s SDK.
An SDK (or software development kit) allows apps to communicate with Adjust’s servers. App developers integrate the SDK into their app’s code, much like if they had a car and a manufacturer gave them a new part for a bit of an upgrade. This creates a line of communication between the app and us through which we can provide attribution data in real time.
Attributing data: attribution windows and the attribution waterfall
Once a user’s installation has been confirmed, an attribution provider begins to look at their past ad engagements and attempt to make a match. It's also important to understand that Adjust won't look at every past engagement - only the ones that fall within the attribution window.
An attribution window (or conversion window) is the period of time in which a publisher can claim that a click or a view led to an install.
For example: if a window of seven days is agreed upon between an advertiser and publisher, and it can be proved that a user interacting with an ad from a publisher installs the app within the window, then that publisher is credited with the install and would receive payment.
Attribution windows are an essential tool for helping advertisers and publishers to understand when a conversion takes place. Often there’s a gap in exposure to an advertisement and an install, such as seeing a game on Facebook while commuting to work in the morning, seemingly forgetting about it, to then remembering to install on the way home from work in the evening. Setting an attribution window creates the ability to include users who are technically brought in from an advertisement, just not directly in the instance of seeing it.
Adjust works backwards, looking for the data point within the attribution window with the most robust information, before heading down the waterfall to find the data point with the least amount of information necessary in order to make a confirmation:
Advertising ID match: first, we look to see if we have any past click ad engagements with the same advertising ID. On iOS, the advertising ID is called an IDFA. The Android equivalent is called a GPS ADID.
- Android referrer: for Android phones, we’ll also check for a match via PlayStore referrer, a unique value our SDK assigns to the specific click. These IDs are not volatile and are as accurate as advertising ID matches.
- Click fingerprint match: if the above data is unavailable, we will look to see if we have any past click engagements that came through with the same IP address and device name; this can be tricky as IP addresses can be dynamic and change rapidly, especially if the user is on the go, which is why for these types of data points we use a much shorter attribution window.
- Impression device match: if this device is unavailable, we will look to see if we have any past impression ad engagements that came through with the same advertising ID.
- No match: if Adjust goes through all of the above checks and has not found a single match, the user is attributed as organic.
Those are the basics of how Adjust goes about attributing data points. Now you know how essential it is to incorporate an attribution provider that follows the user from first engagement to their very last in-app purchase, as well as what kind of data you can gather from that first attributed install.
If you've made it to this paragraph, you'll have realized that there are no cookies, image pixel tags or any of the other standard web tracking methods anywhere to be seen. That's because mobile attribution is completely different from web attribution, and the methods that work for web tracking won't work here. In our next post, we'll be talking about why that is, and what it means for marketers entering the mobile space for the first time.
Want more info on how mobile attribution works? Stay tuned for the next post in our beginner’s guide series, or dive into our new e-book, ‘The essentials of mobile app attribution: A guide from start to finish’.
If you enjoyed this post, you might also be interested in learning about how TV attribution works, or our blog post sharing what you can learn from attribution reports. We also have a useful post devoted to multi-touch attribution modeling.