Anamap Blog
Don't Waste Time, Implement Your Customer Data Platform Correctly The First Time
Data Quality
6/18/2024
Setting up your data plan or tracking plan correctly out-of-the-gate is critical to your company's long-term success with a customer data platform (CDP). A poorly executed tracking plan will lock you into a way of organizing your data that will make it more difficult to analyze the data and after you've already collected a bunch of data it can be challenging to convince the organization to start over with a new tracking plan that better serves the business.
What is a Customer Data Platform?
At their core customer data platforms serve to unify analytics about a customer by tying web analytics, purchase data, email vendor events, or any other off-platform events to a single user in the system. The benefit is that you're able to see a user’s interactions across all your brand touch points and use that data to take key marketing actions. Beyond data unification customer data platforms also provide a means of measuring and enforcing data quality. Segment, ActionIQ, Lytics, and Treasure Data seem to be some of the most popular customer data platforms currently with some newer platforms like mParticle growing quickly. The space is rapidly expanding with new companies and new features being released nearly weekly. All of these platforms have one thing in common which is their ability to help enforce data policies in the form of tracking plans that should be leveraged to improve your company's data quality.
What is Data Quality?
Data Quality is a measure of how consistent your organization's data is. Higher Data Quality means your data has fewer instances where events are missing properties / attributes (or have extra ones they shouldn't) and those attributes have the correct values attached to them. Naturally, low Data Quality means there are more errors and therefore more noise in your data.
Why is Data Quality Important?
The more noise is in your data because of lower data quality the harder it is to get clear insights. The patterns tend to be less clear or require more caveats when the data quality is low. If the quality drops low enough people searching the data for answers may come up with different answers which inevitably errors confidence in the data. Once your company has lost confidence in your analytics data it slows down the pace of decision making because leaders thoroughly question insights or may even dismiss them purely due to trust in the data. If you ever find yourself in this position it will be a long uphill battle to regain that lost confidence.
How can tracking / data plans in CDPs improve data quality?
The tracking plan in your customer data platform, like Segment, is essentially a contract. The contract is between the platforms that serve as your data sources and your CDP. This contract helps ensure that only valid data is collected and stored in the platform. Depending on which customer data platform your organization works with they have different levels of blocking and different bells and whistles to accompany the various kinds of divergence from plan.
What is the right way to setup my data plan in Segment, mParticle, Amplitude, etc?
First, each page template should have its own view type event name. Second, each view type event should contain a property or attribute called event_type (or similar) and set to “view” so that your organization is able to easily calculate page views for an entire user experience. There are two major reasons to use this format for your schema. One: the way that the tracking plans are built can only be as specific as the event name is. In other words, if you have a generic page view event you need to add every conceivable attribute on the event for every page view on your site which means fewer of them can be required. If you use page templates to define the event name, then each page template can have its own unique set of attributes which means your tracking plan can be more specific and therefore more effective. Two: most analytics platforms for visualizing the data are designed around having unique page event names. When looking at the event stream it's easier to see that a user went from the homepage to the catalog page to the product details page instead of having three non-descript page view events in a row that required expanding to see which specific page they were.
Is there an easy way to maintain your data plan / tracking plan?
Yes, Anamap has you covered. Though most CDPs have a UI for managing your plan they are hard to use and aren't tied to any other system. Anamap allows you to easily manage your attributes, events, and views along with their respective relationships which makes the process of managing the data plan easier which means fewer costly errors. Soon we're releasing a feature that allows you to manage your plans entirely in Anamap and sync the updates to your company's customer data platform. You can manage your tracking plan in one place and have it sync everywhere.