Data Validation

Why Does Data Quality Matter?

It is rare to see an executive say “I couldn’t sleep last night because eVar42 has several data issues!” This is unfortunate since, in digital analytics, the “devil is in the details.” So how can you get your executives and your organization to care about data quality? And how can you build data quality into your daily processes?

Common Challenges to Data Quality

Cost. Organizations spend enormous amounts of money on digital analytics. Consider the costs for analytics tools, the analytics team, the developers who populate the data, and the indirect costs of the people who spend time using the data that is collected. All of these costs are wasted if the data they are using is incorrect.

Risk. If the data in your analytics implementation isn’t accurate, you can lose additional money by basing business decisions upon faulty data! 

Resources. Most organizations don’t devote enough resources to data quality, and verifying the quality of data on hundreds of data elements is arduous and time-consuming.

Infrequency. Many organizations only spot-check their data or fix data issues when they are noticed by a business user. Others pay additional money to use 3rd party tools to identify data quality issues.

Trust. But data quality isn’t something that should be left to chance or tackled on a periodic basis. If digital analytics data quality suffers, it tarnishes the reputation of your implementation and your analytics team. If your business stakeholders encounter too many data quality issues, they will stop using your data to inform their business decisions.

Inattention. Compounding the problem is the fact that executives tend to care more about higher-level goals than whether data at the lowest levels is accurate. While it may impact their analytics team or business users, executives don’t feel the direct pain associated with analytics data quality issues.

How Apollo Provides Real-time, Interconnected
Data Quality Validation

Apollo understands and connects all aspects of the analytics implementation. When building the data layer, Apollo uses the selected business requirements to identify the exact data layer elements needed to answer the selected questions. Since it knows what data is expected for each data layer event, if it can understand the structure of the data it is expecting in the data layer, it can validate each data element as it passes through the data layer. Validating the data quality can be done in real-time and be automated.
Let’s illustrate this with an example. Imagine that you work for an eCommerce company and have online orders as part of your analytics implementation. You might have a data layer that contains an Order Placed data layer event like the one shown here.
You can see that one of the data layer elements is “currency” and that it is expected to be a three-digit string of uppercase letters. By carefully defining the parameters of this data layer element, Apollo can “watch” data being passed through the data layer and flag cases where the “currency” element is not a string, not three characters, and/or not uppercase. Any of these scenarios would indicate a data quality issue. These data quality issues can be flagged in debugging consoles or they can be pushed to Apollo itself. Here is an example of a real-time debugging console with several data quality issues.

What This Means for Your Business

Check

Visibility

The analytics team and developers can see real-time data validation issues at the data layer level in debugging consoles or within the Analytics Management System.
Check

Simplicity

Data validation rules can be set up once and then leveraged continuously to find and fix data quality issues as soon as possible. No more manual data quality checks or spot-checking data. You can know every time there is a data quality issue in real-time.
Check

Prioritization

Since business requirements can be marked as “Critical” or “Average,” Apollo can be configured to send different types of alerts based upon the importance of the associated business requirements. For example, a text message may be initiated for a data quality issue impacting a “Critical” business requirement, while an email may suffice for an “Average” business requirement.
Check

Contextualization

Apollo connects data quality issues to business requirements. Executives don’t typically care about individual data elements, but they should care about the business questions that can and cannot be answered by the analytics team or its stakeholders. Since Apollo knows how each data layer element supports each business requirement, it can highlight the exact business requirements that cannot be answered as a result of data quality issues. This means that executives and stakeholders can finally feel the direct impact of data quality issues.

Check

Trust

This can help earn executive support for the time and effort that goes into supporting data quality issues within your analytics implementation.

Ready to take a spin around Apollo?