If data management is the spine that connects all aspects of the data lifecycle, then your tracking plan is the brain that controls it. A well established tracking plan that is connected to your schema registry will become the control center for all the data flowing throughout your business.

Here are 5 principles of data management that should be enforced or controlled by your tracking plan. This checklist is a good starting point for developing your data management strategy.

1. Real Time Data Delivery

While building your cloud based stack you are likely to ensure real time data delivery as a key aspect of your decision making process. As an example you may be working with Snowplow to deliver custom data from your mobile app or website in real time.

If at this stage you are still relying on an excel sheet as your tracking plan, it will be impossible to leverage the benefits of real time data and keep control of the tracking plan, unless they are properly integrated.

It is important that your tracking plan solution is able to integrate directly with your data management platform, ingest your schema registry in real time & have alerts that are also delivered in real time to inform of any potential issues or discrepancies with your data. See here how Snowplow monitor and manage data in real time.

2. Ownership & Stewardship

As with all projects & business processes, clear ownership is always necessary. There are 2 ways to consider the ownership of data - one would be the legal implications and ownership of customer data that a business will acquire, and the other is the person, team or business unit that is responsible for this data within an organisation. Both aspects need to be considered when creating your tracking plan. One particular thing to consider is how your tracking plan integrates with your data architecture.

3. Data Meaning, Quality & Governance

Imagine a very likely scenario where a marketing team at an ecommerce company is busy mapping out some custom event data and simply naming an event:

Order Completed
Order Complete
Checkout Completed
order completed
Checkout_Complete

These events, are all, in theory, tracking the exact same user behaviour in an Ecommerce app. And this list continues - it continues in the event naming convention, the definition of the user behaviour the event represents, and also the properties associated with each event.

A tracking plan should be the tool that defines, centralises and makes visible the meaning of each data structure and associated properties for the organisation. Together, via integration with other tools in your stack, it should publish the schema registries that include the various rules, conventions & other specifics that can be collated in 1 place to ensure data quality visualised as a single source of truth. Through real time alerts, versioning & explicit user interaction required for updating & changing aspects of the data, essential governance is baked in.

Read more about how Snowplow think about structuring event data.

4. Collaboration

This may be easily overlooked, but in any modern day business working with behavioural event data that is at the core of so many business functions, ease of collaboration is essential.

Too often we’ve seen the ever present excel document, aka, tracking plan, aka data dictionary, or some disjointed hybrid of the two floating between teams. The product owner needs to get input from a marketing team that needs specific custom event data for actioning campaigns, while the analytics team has an entirely different requirement, and your development team is left somewhere in between trying to implement everything for everyone.

By the end you’re left with a document that nobody is maintaining, is out of date almost as soon as it’s created, and unable to add anything to data meaning, quality or governance.

5. Maximise Data Use

This point is more of the output of a well designed data management system. After all, there is little point having the cleanest most real time data correctly presented to all functions across a business if it is not actionable.

To this end, a solid tracking plan that can act as the control center, or “brain” of the data management system, is key to the end goal of uncovering well maintained actionable data. Whilst a tracking plan is not the platform or tool that directly relates to execution, it is the tool that should be doing all the hard work at the top of the data stream to ensure little effort required once your data is available downstream in the relevant destinations.