Move from device-based to people-based analytics and personalization

In recent years, brands have been forced to respond to the rapid expansion in the number of devices and platforms customers use to access digital content. Less than a decade ago, a single website and/or web-based application was sufficient for a major brand to deploy its entire library of digital content. Fast-forward a few years, and it’s not uncommon to see comparable content delivered on websites, social networks, mobile apps, and other internet-connected devices. While vast amounts of resources have been spent to align these experiences, in many cases, the same cannot be said for the data that they produce. When it comes to behavior data, you need to create a common taxonomy that is consistent across any device your customers use to interact with you. Further, if you have an existing relationship with the customer, you should make every effort to combine that data into a user-based (as opposed to device-based) profile. Thankfully, the emergence of powerful data platforms like mParticle make it viable to do both, in order to deploy analytics and personalization strategies that speak to the individual not the device user.

Often times, when our team begins to gather data to help clients optimize their digital media, we are forced to contend with disparate data sources using uncommon classification systems and methods. Imagine a brand that uses Google Analytics to track a website visitor who signs up for a “Trial Download.” Later, the same user requests a “Free Trial” on a smartphone app which is tracked using Localytics. We may be talking about the same customer event tracked in multiple vendor databases, with different naming conventions. Now imagine trying to merge that data in order to create a group of unique individuals to whom you will send a personalized message around the time the free trial expires. If you have ever worked with multi-platform behavior data, you understand how challenging that would be. Now imagine the same scenario, where data flows into a single source, bound by a single customer identifier, tied to a common “Free_Trial_Event.” The analytics platform and device become irrelevant. We now have a unified data source that allows us to transmit ‘people-based’ data wherever and whenever we need it. This illustrates the necessity of a data platform like mParticle in a multi-device customer journey.

When planning for a data platform implementation, it’s vital to include every stakeholder group across product, analytics, and marketing teams. When an analyst defines data requirements, it’s often behind a lens of gathering user behavior insights, while a marketer may be focused on media performance and attribution. Each team has unique needs when it comes to the user attributes, content naming, and event data that must be captured. The data should be intuitively named for use in various personalized communications like email and notifications. In addition, many of our clients use multiple analytics vendors for reporting and analysis, so it also must be compliant across vendor best-practices. A seemingly small issue, like using upper/lower-case characters, can create major issues later on. We recommend being as explicit as possible when it comes to formatting data, with ongoing compliance enforcement. The goal is to keep your data clean, complete, and compliant, therefore, thorough stakeholder interviews are a necessity.

After we have fully defined and implemented a data plan across all device platforms, (e.g., ‘ios,’ ‘android,’ ‘web,’ etc.), it’s time to focus on which vendors will be receiving data, how much they need, and exactly how it will be transmitted. By this point, the heavy lifting has been done. mParticle provides a robust inventory of vendor integrations that allow you to simply drag and drop the data inputs and vendor outputs into a “connection.” For instance, in a basic connection, an attribution vendor like Kochava may only need to capture native app installs and a single key event, such as a subscription purchase. With a few clicks, all the data you need will start to flow in real-time between your mParticle and Kochava apps. In a more complex connection, say a push notification vendor who sends targeted messages, the connection might require various user attributes, or a grouping of users who have previously performed a series of actions. In this case, you simply toggle the events and attributes you want to send and create business rules for sending them. With a few exceptions, mParticle’s integrations transmit data via server-to-server connections. When you couple that with the elimination of most vendor-specific libraries, we often see significant performance improvements in our applications.

As digital content delivery continues to fragment, with multiple vendors vying for data, it makes sense to focus first on standards and unification. Once you are able to compile a unified data profile of an individual, it becomes possible to understand the entire customer journey. Ultimately, this is the key to personalized and appropriate communication at each step along the way. A data platform such as mParticle is a vital tool to facilitate that process.