This is the third installment of mParticle’s “Connected Marketing Stack” playbook.
As everyone knows, there are a multitude of technology platforms and tools in the modern marketing stack and most offer APIs to communicate with one another. That’s the good news.
The problem is that they do not all connect in the same way or equally well, and there are no standards. Meaning that if you have 20 platforms and tools in your stack (and many marketers have far more) that could require as many as 400 integrations to set up and maintain!
That’s why marketers need a dedicated connectivity layer. Whether you decide to build one yourself or license a solution like mParticle’s, here’s a five-point checklist of what it needs to do:
- Collectively Exhaustive: Your connectivity layer should be “fit for purpose” for powering your entire stack, not just a component of it. Many digital marketing hubs built for CRM and web execution lack the means to collect the full spectrum of app data from mobile and connected devices, putting the onus on the organization to figure that part out. Conversely, many best-in-breed mobile solutions are deficient in other areas, such as data ingesting data from the mainstream marketing cloud providers and/or data warehouses. A good connectivity layer needs to support all the data types you require to power all the tools in your stack.
- Open: Your connectivity layer should be partner and channel agnostic. It should be able to expose any and all data you need from it to everywhere you need that data to go. This also means that its API endpoints should be well documented so that developers can write to them should there be a need to add services that are not already configured.
- Low Maintenance: Your connectivity layer should eliminate, or at least minimize significantly, the overhead associated with maintaining third-party APIs. Because as APIs change, which they often do, you will need to update your systems, you want to ensure that those changes are a) well telegraphed and also b) don’t create any downstream challenges. This is particularly true of APIs that rely on mobile app SDKs since they will require an app store update pushed to all your end users’ devices. That’s why, whenever possible, you’ll want to opt for a single point of data collection and an abstracted approach to data connections.
- Flexible: Your connectivity layer should be able to “translate” in order to support all of the various schemas utilized by each of the endpoints across your marketing stack. For example, data may need to be mapped to meet a certain vendors’ fixed taxonomies or data models. Or, while some systems may accept a real-time stream, others may impose throttling limited such that you may need to batch data to them. Your data platform must be flexible enough to accommodate these requirements from all relevant “consuming” APIs.
- Secure: Your connectivity layers needs to be secure and reliable from end to end. This is not a trivial point, as data leakage is real. Data must be encrypted along the entire data journey, from when it’s collected to when it’s at rest to when it’s transmitted downstream. There should also be live monitoring capabilities to troubleshoot bugs before they become an issue, error alerting so problems don’t stay hidden for long, and the ability to “replay” backup data at any point, if necessary.
Master these fundamentals and you’ll be able to connect and control your customer data like a boss!
David Spitz is CMO of mParticle. Follow him on Twitter (@david_spitz). Want to learn more? Check out mParticle CEO Michael Katz’s Mobile Apps Unlocked 2016 keynote “How to Create a Winning Mobile App Data Strategy” here.