The mobile data opportunity
Mobile isn’t just another channel or ad format, mobile is a completely alternate view into the consumer. The consensus is that first, over the next few years all users will be mobile almost all the time, and we’ll see the continued unbundling of web experiences into app centric experiences, extending into wearables, and automotive UI’s, and Smart TV’s. At the heart of unlocking the opportunity is the user interaction data that is created. Because of the size of the opportunity, it’s important to understand the differences in data between web and app environments. And while many are quick to point out that lack of cookies is the reason that capturing app data is different, that’s very much an incomplete answer.
To begin to understand the difference, it’s important to first review the deployment model differences. Over the past fifteen years there has been a massive shift toward agile product management and ultimately rapid software deployment via the web, which further accelerated with the adoption of cloud based services. Native apps however, are software products that are typically deployed in fixed release cycles, sometimes quite infrequently. This shift initially represented a step back in terms of speed and flexibility from web development, which created both a challenge as well as an opportunity. Point solutions came to market to address some of these challenges, for example TestFlight for testing and development as well as Crashlytics for crash reporting then Analytics etc. Pretty soon the service provider ecosystem started to take shape to help developers run their app business with the speed and flexibility they had become accustomed to on the web.
Getting into the differences in data, mobile app data is highly dimensional compared to web data. There is considerable fragmentation across carrier, device manufacturer and model, operating system including OS version, app software version and other technographic features. Constructing segmentation using heterogeneous data structures such as app events, exceptions, and push notification receipts across various hardware and software dimensions and again across a geo-spatial plane will push any data science team to their limits. Lets look at a few of the other differences:
One significant difference is crash data. Crashes (exceptions) are events that represent a really rich CRM opportunity and ultimately an activation point, especially for new and high LTV users. When a user, especially a new user, has a poor experience they typically will not return to the app anytime soon. Being able to specifically interact with these users outside of the core app represents an immediate and critical opportunity to engage with users in a way that allows app owners to turn a negative experience into a positive one.
On the subject of geo-spatial complexity within mobile, it’s important to look at not only what a user does but also where that action happened and potentially even who they were with. In mobile, context isn’t just about content, which means that data models built exclusively around page level data must be re-imagined. The reason mobile is such a massive opportunity is that it is the glue between the digital and physical world. As mentioned, context extends beyond the device into the physical world, the people you’re with and places you’re at means that the same behaviors in different contexts can mean completely different things. This has major implication on certain tactics such as re-targeting especially since ads usually don’t remind people to use certain types of apps, whereas environmental context usually does.
Device telemetry data is unique to mobile as well. Data points such as velocity, acceleration, and position are pretty much non-existent in the PC world, yet they represent a massive opportunity in the mobile world. Things like iBeacon and geo-fencing are creating significant opportunities to help businesses not only understand who their consumers are but where they are and engage with them in new and exciting ways.
Browserless environments and interactive messaging also represent new challenges and opportunities. Being able to send a push notification to a user when they are not browsing the web, using an app, or even logged into their phone is a significant shift from any and all previous engagement techniques. With the release of iOS 8 now imminent, the possibilities here are only becoming even more interactive and exciting.
Because of the significant differences mentioned above, data solutions such as DMPs and Tag Management solution which were built for web-based environments simply don’t work in app environments. Marketers and publishers need to work with systems purpose-built to handle the unique complexities of software-based environments where things like cookies and pixels don't matter. Ad hoc, pixel-based integrations are no longer relevant in an native apps and neither are the rules of (re-)engagement.
In parting, while we can all acknowledge that data activation is fundamental to creating value at scale, legacy solutions which were built for web environments and have only integrated into other web solutions will see their importance diminish over time. There is a flourishing ecosystem of innovative tools and services built for native apps that help you stay close to your customer which should be connected through a common fiber. The ecosystem is evolving fast and the time is now for a data layer that is compatible in these types of environments.