How To Make Sure Your Customer Data Platform is…

Data at rest will stay at rest, and data in silos can become trapped and opportunity lost. That’s why a number of companies have sprung up to help marketers, developers, and publishers simplify the process of collecting and connecting data to various analytics and marketing services. Rather than embedding a number of client side SDKs directly into the app, which require extensive engineering commitments each time, the data hub serves as an abstraction layer and provide economies of scale from a single engineering initiative. Implement a single collection point to consolidate data capture and streamline data flow to improve agility.

Creating control around first party data democratizes value across the organization. For developers, it simplifies the data collection process while it improves speed and flexibility for marketing and analytics teams. To accelerate data activation, API integrations allow customer data to be distributed without any additional engineering work, but not all integrations are created equal.

For a data hub to create meaningful value for the business, simplifying the collection process is a good start but it’s only the beginning. The real value is created once the data can become activated across multiple services seamlessly. As such, integrations must extend beyond minimum viable connectivity and aim to support as close to full feature parity as possible. When evaluating a data hub the number of supported integrations matter because discovery of new services is important, but integrations are not just about the sheer quantity; quality and relevance are what create the most value. In this post, my aim is to provide deeper insight into how to make an informed decision about investing in a customer data platform.

The Data

Lets first start by building a deeper understanding of the data since this is what ultimately enables integrations. While there are many different types of data models, every marketing and analytics service is powered by data that answers who, what, and where. And at the heart of any integration, there needs to be consistency in the data that is captured and ultimately distributed. But when choosing a customer data platform for mobile, it’s critical to not become mislead by data models that are rooted to a web paradigm. Apps generate data collection opportunities that fall outside the limits of a web-centric framework with data such as push tokens, exception handling, and device telemetry data, among others. Without accounting for these native data types, the types of integrations provided by a data platform will also be limited.

Additionally, one quick check to determine if the platform serves your mobile needs is to see if there are references to “pages” as opposed to “screens” in the docs. This may not seem like a big deal, but if the platform references “pages” for something that happens in apps this is a telltale sign that the platform wasn’t purpose built to solve for mobile data challenges. For a customer data platform to fully meet the needs of both the mobile marketer and the mobile developer, it needs to be able to capture all of the available data types core to app businesses, and probably should have been purpose built to do so.


As I mentioned, not all integrations are created equal and what determines quality is the usefulness of the integration. When it comes time to implement a data hub, it’s important to dig a level deeper and understand what each of the integrations offer, what data can be sent, and what functionality is supported. Then map that functionality to the problem you’re trying to solve. If everything maps properly and you are not sacrificing partner functionality in the process, that is a high quality integration.

Besides Analytics services, which are often very straightforward integrations, the most often implemented tools which consume first party data are Marketing Automation tools. The leading Marketing Automation tools work by providing app marketers with the ability to engage users inside and outside the app using data to deliver targeted push notifications, emails, in app messages, etc. Simply sending event-level data to one of these solutions via a data hub will certainly streamline the tagging process but it ignores the actual needs of the mobile marketer in terms of activation and improving speed to market.

Additionally, solving for the “who, what, where” model of data misses the most important use case around activating that data. Without transferring push tokens, partner functionality is limited and requires a separate partner SDK integration to activate the data, which doesn’t actually help accelerate development or marketing initiatives. The key in creating a useful solution for marketers is not only streamlining data flow but also ensuring usability is not sacrificed or eliminated.

Lets look at an example from Segment:

Here you can see the type of integration supported by Segment. Appboy is a powerful customer engagement platform and Segment’s integration allows apps to send user IDs, custom events, and group identities to them. It is important to get data into a partner like Appboy’s system but is a far cry from supporting Appboy’s core user engagement capabilities. So from a developer standpoint, the collection process may be simplified but they would still have to integrate the Appboy SDK to activate full feature functionality. Developers beware: if the ability to streamline data doesn’t also improve or at least maintain usability of the service, you are not only missing the core business use case but you are creating more work for yourselves in the long run.

Alternatively, the ideal integration offers full feature support of the 3rd party vendor. This not only allows app developers to streamline the tagging process but it allows app marketers seamless activation without any additional code changes, or delay in resubmission to the app store for approval. Again, app developers not only need to streamline data collection but also empower app marketers to realize the potential of their data in real time.


While lots of integrations are great, offering the ones that matter is even better. A major difference between mParticle and Segment is our commitment to helping app marketers connect with the services that matter to them. For example, the data management platform (DMP) has become a critical piece of infrastructure to serious digital marketers, allowing them to gain a multi-dimensional view of the consumer across web, search, email, and other channels. Having integrated with industry leaders such as Tapad, Lotame, Krux, and BlueKai, mParticle ensures that any omni-channel marketer can always get all their data into the platforms that matter to them.

Additionally, key measurement vendors such as Nielsen and comScore are critical to app marketers and publishers. They provide currency both internally and externally to validate a number of important data points. Without these types of measurement integrations, how much time is actually being saved? Make sure you’re not being tricked into an 80/20 solution. 80/20 data solutions provide coverage across 80% of the type of data required to be collected but only 20% of the actual use cases where value is created. A customer data platform should aim to balance simplicity and usefulness from both a collection as well as an activation standpoint.


Ultimately, data must be easily activated; as it’s the difference between potential energy and kinetic energy. Anyone can create integrations which offer minimum viable connectivity, aggregate lots of logos, and market quantity as a cure-all. The real difference is quality and relevance. To provide quality and relevant integrations that empower the business, it starts with capturing all of the data required to power these integrations. Some platforms are great at helping simplify the data collection piece, but they miss the mark on making an impact on the business since they cannot offer feature-complete integrations due to limitations in their data model. Using solutions that were not designed to capture many mobile-only data types will only create more work in the long run for developers as well as marketers. At mParticle, we are committed to helping developers be more efficient and marketers more effective.

Show your support

Clapping shows how much you appreciated Michael Katz’s story.