Mobile advertising is all set to reach $247.4 billion in 2020 globally. It’s becoming increasingly important for app marketers and product managers to understand who their users really are. Marketers need to be able to identify which platforms, sites, and ad networks are generating installs and where their resources are getting wasted. That’s where mobile attribution comes in.
Mobile attribution is the process of matching potential users’ data with the key actions in their journey towards becoming a paying user. It helps app marketers and product managers to improve app performance by applying deducted data to their marketing campaigns.
In this article, we’re talking about one of the most popular and widely used methods to track mobile app installs – device fingerprinting.
What is mobile device fingerprinting?
Fingerprinting, also known as device recognition, is an attribution method in mobile advertising. It uses publicly available parameters such as the device name, type, OS version, IP, carrier, etc. to identify a successful app install and form a digital fingerprint ID that matches defined device attributes.
Mobile device fingerprinting is predominantly used on iOS due to the lack of referral data from the App Store. In some cases, it is also used on Android when Google Play Referrer data is unavailable.
Why is mobile device fingerprinting important?
There are several ways to track which marketing campaigns directly attribute a click to an app install based on a device’s ID (such as IDFA or Google Advertising ID). However, the device ID is not always available. Sometimes it’s because a user turns on limited ad tracking in their settings or the click comes from the mobile web or it is not passed by an ad network.
When deterministic identifiers like device ids are not available, the only alternative available is mobile device fingerprinting. Although probabilistic in nature, it enables app marketers to understand their users better, and also track the performance of their campaigns closely. Here are a few reasons why the attribution method has become a popular choice:
1. Not all advertisers support mobile identifiers
As mentioned above not all ad networks pass along the device identifiers. If an attribution technology does not fallback to mobile device fingerprinting then app marketers lose out on critical attribution data. In many cases, they won’t know the complete user journey and the performance of their marketing campaigns.
2. Users can choose to turn off tracking
Even if the advertising network supports the mobile attribution identifiers you want to use, sometimes it’s your users who create the data gap. With data privacy becoming an increasing concern for many, you’ll find some users turning off ID tracking on their devices.
In this case, too, app marketers and managers end up struggling with user data. They can’t identify the key data that defines their users or tells them more about their behavior, to optimize campaigns.
3. Longer time gaps make data redundant
If the time gap between a click and an install is too long, there is a chance for the user to change the settings on their device: switch internet connections (WIFI to 3G for example), or simply use other devices matching the same fingerprint. This results in most attribution methods tracking new identifiers, creating a fresh user record every time. For example, when the mobile user’s IP address keeps changing with their location, fingerprint matching will get weaker and result in mismatch!
For most app marketing and advertising campaigns, the majority of clicks, installs and the first use of the app happens in the first 1-2 hours, in most cases in minutes. Ironically, setting such short attribution time windows sometimes introduces inaccuracy for the exact opposite reason. In such instances attribution providers will misattribute users as organic installs when they convert outside of this attribution window.
How does mobile device fingerprinting work?
App install attribution with device fingerprinting begins from the moment a user clicks on a URL. The JavaScript that is placed as a tracker by the app starts to collect all the data around the user, like:
- IP address
- Platform
- Device brand
- Device model
- Device carrier
- OS name
- OS version
- User-agent
- Timestamp
- Screen resolution
- Browser version
- List of plugins
- Language
A device fingerprinting provider then combines the above data with browser cookies and other identifiers to assign a unique fingerprint. This data can then be further enriched by correlating it with other fingerprints and identifying similarities between users coming from one source. With enough data, sometimes an identity can also be associated with multiple devices.
Let’s take a simple example to explain how the device fingerprinting works.
Every time a user clicks a URL, they leave behind a few clues about what device they are using and the intent. These clues are collected by the attribution provider to create a profile (non PII) for the device known as the device’s fingerprint. Users are then redirected to the app store to install that app. Once they launch the app post install the same information is captured again and matched against the previous fingerprint. The attribution provider at this point will determine the accuracy of the match and enhance it with any historical matching data available. When a match is successful the install is attributed to that device.
Different types of software and online services already use a form of device fingerprinting. Some of them are not aware of that because having Google Analytics installed on their site already enables detecting users’ device, location, OS and many other parameters.
How reliable is mobile device fingerprinting attribution?
Mobile fingerprinting can never be 100% accurate.
The inaccuracy occurs due to a few identifiers that are available for device fingerprinting. Clicks originating on mobile web only have publicly available data from the browser’s user-agent and the IP address.
Assume there are two users connected to the same public wifi network i.e. having the same public IP address. Additionally, they have the same device type, model, OS, and OS version. Mobile device fingerprinting labels these two users identical creating a ‘false attribution’ or no attribution at all.
Here’s what it means.
User A could have discovered your app on the app store and then directly downloaded it.
User B could have seen an advertisement, clicked on it and then downloaded your app.
In both cases, user journeys are very different. Being marked identical doesn’t let app marketers optimize this journey for higher conversions. When they can’t identify where the users came from, what content they consumed or which ad they were exposed to, there is no way to optimize the app install campaign.
To combat the above inaccuracy, attribution providers set defined time windows for matching fingerprints. But this brings additional inaccuracy of the opposite nature – providers will attribute users to ‘organic’ installs simply for converting outside the defined time window; which is usually less than 24 hours and in most cases just 2 hours.
In addition to the inaccuracies, fingerprinting requires large amounts of storage. Unlike cookies, fingerprints of users are not distributed back to the browser. That makes calculating the IDs (or the hash) and storing all the data associated with it more complicated.
Simply put, there are always inaccuracies in mobile device fingerprinting attribution and it can be overwhelming. Unless you’re using a robust tracking and attribution solution.
Tracking and attribution with GetSocial Smart Links
You’re running various campaigns on multiple channels to drive more mobile app installs. To track your efforts, you’re also making use of different trackable URLs for each platform – be it the one-off links to share on social media, for the search ads or your in-app referral campaign.
But how do you make sure you’re able to track and attribute the right user data through these links?
How do you make sure that the inaccuracies in device fingerprinting attribution are minimized?
GetSocial Smart Links use a proprietary device fingerprinting technology that provides complete tracking and attribution. No matter what platform or advertising network you’re using them on, you get actionable insights into your users and the app’s performance.
Here’s how GetSocial Smart Links work:
Why use GetSocial Smart Links for device fingerprinting?
Smart Links can be easily used across different channels, platforms, and campaigns. Some of the features that make them a popular choice amongst top-grossing apps like Sonic Forces: Speed Battle:
1. Contextual deep linking and deferred deep linking
While it works similar to deep linking, contextual deep links add an additional functionality of storing information about where the link is being used – campaign name, channel, medium and other custom data provided by you.
This gives app marketers the ability to push metadata through the App Stores right into the app and to create a personalized user journey for those who interact with the campaign link.
2. Webhooks
One thing that continues to change is a user’s preferences and behavior. This impacts how they interact with your app install campaigns, their journey to an app install and how they use the app.
Webhooks provide your real-time event data for every app install event. It also adds all the contextual information you need to optimize your campaigns.
3. Tracking and attribution
Smart Links are not just easy to use and share, but also make it easy for you to track campaign performance. Apart from giving you the ability to track which campaigns and platforms are driving more installs, Smart Links also provide insights into the app’s k-factor, cycle time, conversion through invite funnel in a referral campaign, retention uplift and more.
The links also automatically detect the device of the user and then redirect them to the appropriate store – iOS users to the App Store, Android users to Google Play and desktop users to a landing page that gives them an option to receive an app download link on their phone.
Mobile app install tracking, simplified.
4. Retargeting
You can add your Google Tag Manager ID to the Smart Links. So when users click on the links but don’t convert to an app install, you can use retargeting campaigns to turn those visitors into successful app installs.
It’s similar to how you show interest in a product on an online store. But when you leave without purchasing it, you start seeing ads promoting the same product on social media and other websites.
5. Frictionless invites
Word of mouth is the fastest way for an app to grow. 92% of users tend to trust recommendations from their friends and family, over advertisements.
Smart Links can easily be customized to match your branding and are included in Smart Invites. Users can send invites over social media and all popular chat apps like WhatsApp, Messenger, Twitter, Facebook, and others. But more importantly, they can do so without having to authenticate with your app or with any of the supported invite channels. Read more about it here.
Do you need mobile device fingerprinting for attribution?
With the number of apps in the App Store and Google Play increasing by the day, it is becoming challenging for app marketers to get attention to their app – let alone, drive more installs.
The cost per impression and acquisition are on the rise too. So if you’re not targeting the right users, you’re only losing more resources by the day. That’s where mobile device fingerprinting comes with GetSocial Smart Links comes into play.
Getting the tracking and attribution right, you can keep your app install campaigns optimized at all times.
