When your data quality is top-notch, all teams benefit. Product teams can iterate faster and build immersive user experiences with confidence. Analytics teams can build queries without heavy workarounds and inform cross-functional decisions faster than ever. And marketing teams can inspire user action and improve advertising efforts by personalizing messaging according to user behaviors and traits.
Unfortunately, finding data mistakes is far too common for most companies. Whether your
Order Completed event was accidentally implemented as
OrdreCompleted or your
products property was coded as a string instead of an array, you likely have spent time cleaning up your data set just to make it useful. Rest assured, you're not alone: 83 percent of companies have a dirty secret … dirty data.*
Getting your organization to a state of high-quality data that all stakeholders trust takes a combination of alignment, validation, and enforcement. In this lesson, we'll share what it takes to collect high-quality data at scale.
Companies that take data quality seriously typically create implementation specs or tracking plans to align their business objectives with the metrics and events they track.
A tracking plan is a document you can use to standardize customer data collection across your app and website. It serves as both a project management tool as well as a reference document, and generally contains three key pieces of information:
Once the above items are written into the tracking plan, the product engineer has what they need to add the events into the right places in the code base.
Any time you make changes to your code, it's important to QA before shipping to production. Instrumenting events is no different! That's because a single tracking error on a business-critical event, like
Lead Captured, can cost your business hundreds of thousands of dollars.
Once your tracking plan is agreed upon, set, and implemented, you'll want to make sure dirty data stays out of the picture. To do so, you'll need a method for validating new tracking events making their way to your codebase.
Any new event that does not match your tracking plan should be flagged and fixed before making it to your production app.
If you want to really take data quality to the next level, you need a way to ensure bad data never hits your marketing and analytics tools. But that's easier said than done.
One way to manage enforcement is to create a "data governance council" to manage the entire process. The team's responsibilities should include:
If you're already using Segment, you can use Protocols to automate enforcement (and alignment and validation, too!) Once you've solidified your tracking plan and implementation, you can configure your settings to automatically block any data that doesn't adhere to your spec. That way, only planned and approved data will make it to your marketing and analytics tools.
After working with thousands of companies to configure their tracking, we launched a product called Protocols to address the biggest challenges to achieving data quality:
You can learn more about Protocols or request a demo to see it in action here.