The mobile industry is a data-driven business. App owners optimize their apps’ performance and design based on behavioral data of their users. And they optimize marketing campaigns based on attribution data. To gather the latter, they need an attribution tracking tool.
What is Mobile App Attribution Tracking?
Mobile attribution tracking is the process of following a user from seeing an ad via clicking it to the install of an app and beyond. Its goal is to attribute installs and post-install events to the publisher of the ad.
Attribution tracking is an essential component in Mobile Marketing and it is a must-have for doing user acquisition. Read on to learn why.
Why is Attribution Tracking important?
Attribution tracking enables app owners to understand how and where users discover their apps. Attribution data connects ad impressions and ad clicks to app installs, but also to users’ post-install actions. So app owners are able to follow users from their first contact with the app to their last in-app conversion.
Based on this knowledge, they can commit their marketing efforts and budgets to those traffic sources that deliver the most users or the best quality users. So attribution tracking is the basis for optimizing performance campaigns and making the best out of every marketing dollar. As it allows to compare different traffic sources based on standardized metrics, it also helps identify fraudulent traffic.
Besides, attribution tracking is necessary for proper billing. Without it, ad networks do not know how many installs to claim, and thus they cannot bill the advertiser accordingly.
What is the Difference between Attribution Tracking and App Analytics?
You might wonder why you need an attribution tracking tool if you already use Google Analytics or a similar app analytics tool. At first glance, it seems that these tools provide the same data. But actually there are some differences between attribution tracking and app analytics.
App analytics focuses on users’ in-app behavior. It gives app owners in-deep insights about user journeys and their actions, so app owners can identify usage trends, but also problems and weaknesses of the app.
But it unfolds its power only after the install. In some cases, it is possible to attribute a user or a user session to a traffic source. But for most traffic sources, Google Analytics will wrongly count organic installs because it cannot connect them to interactions with an ad.
In contrast, attribution tracking can make this connection. By using tracking links, the origin of an install can be tracked properly. Thus, it is more reliable in terms of attribution.
Attribution tracking providers focus on analyzing user behavior before the install. It enables app owners to understand where users came from and how marketing campaigns perform. Most attribution tools also provide anti-fraud mechanics that help advertisers find and exclude fraudulent traffic sources.
What are the Requirements for Attribution Tracking?
Proper attribution tracking is a sophisticated process that requires a technical setup. A couple of companies in the mobile industry have specialized in app attribution tracking and built integrations with all major ad networks. Partnering with them releases app owners from the necessity to create their own attribution mechanics. Instead, they can sign up for the services of an attribution tracking provider and integrate their technology.
If you want to integrate attribution analytics into your app as well, you need to do three things:
Embed Tracking Links into Your Ads
First, you must embed tracking links into your ads. You can create these links on the tracking provider’s platform. Each link is unique, so it is connected to only one publisher. By adding additional parameters, you can also make it unique for single campaigns and even for specific ads.
Implement the Tracking SDK into Your App
The second requirement for attribution tracking is an SDK that you need to integrate into your app. This SDK is the interface between your app and the server of your tracking provider. It reports app opens and other in-app events that your users perform, given you defined them upfront.
Set Up Postback URLs
Finally, you need to make the tracking provider’s data available to the ad networks you use to promote your app. To do so you must implement a postback URL into the system of your tracking provider. Ask your user acquisition partner for this URL. When a click, install, or post-install events happens, the tracking server sends information via the postback URL to the ad network’s server.
The big tracking providers have very deep integrations with the major ad networks like Google Ads or Facebook, so setting up postback URLs is not necessary. However, for smaller networks and agencies, a manual setup might be mandatory. Double-check with the representatives of your tracking partner and ad network, if you are unsure about it.
How does Attribution Tracking work?
Now that we know the technical requirements for attribution tracking, let us have a look at the process in detail.
When a user clicks an ad, the tracking link sends him to the server of the tracking provider that collects some data about the user’s device. The most important information is the device ID: the IDFA for iOS devices or the GAID for Android devices. But it stores more data, for instance:
- device brand and model
- OS version that is installed
- IP and location
- the exact date and time when the link was clicked
- the device language
- the name of the ad publisher
After collecting this data, the user will be redirected to the product page of the promoted app in the app store. This complete process happens in split seconds, so the user barely notices it.
When the user opens the app for the first time, the SDK collects the same device information that the tracking link gathered before. It sends this install data to the tracking server, which compares it to all data from ad clicks that have been collected so far. When the server finds an ad click from the same device, it attributes the install to the ad publisher who delivered this click.
It is important to understand that clicking the download button in the app store does NOT suffice for attribution tracking. No MMP has access to information about app downloads, so they cannot track them. The user must open the app after downloading and installing it, so the SDK can communicate with the tracking server.
Post-Install Event Tracking
A very similar process takes place when the user performs a predefined post-install action such as registering an account, completing the tutorial, or making a purchase. When doing so, the tracking SDK sends information about the post-install event including device data to the tracking server. The server compares the event data to install data that was stored before. In the case of a match, the tracking provider attributes the event to the ad publisher who was responsible for the install.
Throughout the mobile industry, ad networks use different models for attribution. The attribution model defines which ad publisher will receive the reward (the CPI) for delivering the install. All models have specific advantages and disadvantages.
The most common way to attribute installs to ad publishers is to use the last-click attribution model. In this model, the ad publisher who delivered the last click before the install gets the compensation for the install. It does not matter how many ads the user has seen or clicked before, only his last interaction matters.
Using the last click as the criterion for attribution is a reasonable approach. But this approach is prone to fraud because criminals have developed multiple ways to steal attributions by faking last clicks. Nevertheless, last-click attribution is the standard in the industry.
First-click attribution is the opposite of last-click attribution. In this case, only the first contact of a user with an ad is relevant and the publishers who generated it will receive the reward.
It makes sense to use first-click attribution when the main purpose of a campaign is to raise brand awareness. In the mobile ad industry, it is an uncommon model though, because advertisers usually focus on traffic quality over quantity.
Depending on the tracking partner and the user acquisition partners, you can also attribute installs based on ad impressions. In this case, the publisher who is responsible for the last view before an install gets the reward for this install.
The assumption behind view-through attribution is that the user remembered the ad and headed to the app store to install it later. It is arguable whether this is a reasonable assumption, because as a trigger for an install, an impression is less reliable than a click, anyways. Thus, most ad networks use this attribution model only as a fallback when a user saw an ad but did not click it.
Another model that gets more and more attention in the mobile ad industry is the multitouch attribution model. In this model, all user contacts with ads are measured. So you as an advertiser get a complete picture of how the user learned about your app, but also about how many contacts and interactions it needed to pursue him of downloading.
The graphic above shows a multitouch user journey. When using the last-click attribution model, only publisher C would get credit for the install because he delivered the last click before the install.
But multitouch attribution recognizes publisher A and B as well because the user interacted with their ads. However, there are different ways to distribute the reward for the install:
- Linear attribution: All publishers who were involved receive the same share of the reward. In our examples, the networks A, B, and C would receive 33% of the reward each.
- Position-based attribution: Publishers who are responsible for the first and the last interaction are rewarded. Clicks in between are not. So publishers A and C would split the reward and get 50% each.
- Time decay attribution: The publisher who delivered the last click gets the biggest share of the reward. Other publishers get smaller shares. The bigger the decay between the interaction and the install, the smaller the share will be. In our example, the reward would be split, so publisher C might receive 50% while B gets 30% and A 20%.
Multitouch attribution gives advertisers more insights about users’ journeys to an app download. It also might be considered fairer because the contribution of ad publishers is rewarded even if they are not responsible for the last click before an install. But it also opens the door to fraudsters, because infiltrating a multitouch journey with an interaction to steal a share of the reward is easier than to capture a last-click attribution.
What is the Attribution Window?
When setting up the tracking for a specific campaign, you have to define the attribution window. That is the maximum period of time between a click and an install that still results in a successful attribution. For last-click attribution, a typical attribution window is seven days. This means that installs that happen up to seven days after an ad click will be attributed to the ad publisher. But if it takes eight or more days until the install happens, it will not be attributed and be counted as an organic install instead.
You can set an individual attribution window for view-through attribution. And this is absolutely reasonable as the connection between an impression and an install is less obvious than the connection between a click and an install. Thus, the view attribution window should be significantly shorter than the click attribution window. Typically, 1 to 24 hours are fine.
No matter which attribution model you prefer, you can use different attribution windows for different user acquisition partners. It makes sense to set shorter windows for new partners and increase the length once you have build trust with them. Most of the big players have standard windows though that you cannot alter.
The figure below visualizes the concept of the attribution window. In the example, the window is 5 days. User A installs 5 days after clicking the ad, so he is attributed to the ad publisher. User B needs 7 days to do so. As this period is longer than the attribution window, the user is not attributed and counted as an organic user.
What is Fingerprinting?
Fingerprinting is an alternative tracking method for cases when no device ID is available.
As mentioned before, unique device IDs like the IDFA or the GAID are necessary for attribution tracking. But in some cases, users disable their device IDs, so tracking providers have no chance to collect them.
Fingerprinting is a fallback method to distinguish these devices nevertheless. For every click, install, and in-app event, it creates a fingerprint that contains characteristics like the device brand and model, its language, and the IP. It then compares fingerprints to find matches and attribute installs to the right ad publishers.
The upside of fingerprinting is that it is compliant with data privacy regulations because it does not contain any personal data. Its downside is that it is not 100% reliable. The bigger the gap between the click and the install, the higher is the chance for a misattribution. For that reason, the attribution window for fingerprinting usually is significantly shorter than for device ID tracking.
Attribution tracking is a must-have for anyone who runs an app business. It is absolutely necessary if you do not want to limit your ability to run user acquisition campaigns. And it is also an extremely helpful tool for optimizing these campaigns and identifying ad fraud.
Be aware that attribution tools are premium tools, and providers usually bill their clients based on the number of active users. So the more successful your app gets, the more will you have to pay for attribution tracking. Nevertheless, this is probably one of the best investments you can make for your app, and you should not shy away from attribution tracking because of the cost.