With Data Quality being part of the Tag Management Maturity Model we thought it wise to look a little deeper. When we, at Tagticians, started to reflect on which tag management tools we offer consultancy services for it surprised us that only 2 of the 4 offered any form of built-in Data Quality checks.
In the many years that we have been working in the tag management industry, data quality has been rarely spoken of. The assumption is often made that when a tag is placed on a website, the data passed through will always be accurate and error-free. From experience we can tell you that this is rarely the case.
What are some causes of poor data quality?
With regards to data collection using tag management tools most errors we see today are caused by front-end changes. These errors will lead to the collection of inaccurate and inconsist data. Without a data governance in place, it is challenging to resolve the issues and prevent these errors from happening again.
Here are some examples of potential front-end issues that can negatively influence your data quality:
- Browser bugs result in issues
- HTML Stuctural changes
- Order of code changed on page
- Removal of code dependencies
- Duplication of tags/code
- Javascript library updates
- New javascript code conflicts with a global variable
(source: Tealium)
Front-end errors at least have something in common. Errors occur prior to the data being processed, so preventing poor data reaching your data silo’s is possible.
Checking Data Quality prior to processing
During our checks, it turned out that Segment and Harvest (still in Beta) developed by Grain were the only data collection solutions that offered a data quality checking feature. Segment harnasses its Protocols functionality to achieve better data quality, and Harvest uses Data Cleaning. Both of these features allow for data quality checks prior to processing, in other words, when the data has been collected and prior to processing it for delivery to destination sources.
Segment Protocols Flow
This level of data quality checks has many benefits. For one, it will help prevent your production data from being polluted. Performing analysis on untrustworthy data is a thorn in the side of your business insights. You analytical resources are better spent discovering insights from data rather than cleaning it.
Our recommendation to any company wanting to level up in the Tag Management Maturity Model is to start work on performing data quality checks at the data collection source. In actual fact, defining the data (what, when, where and in which format) that needs to be tracked in your tag management implementation plan, should be the root of all your data quality improvement efforts.
Post-processing audit methods
We would like to point out that there are a handful of tools on the market that you can use to check your data quality. However, many of these tools don’t perform these tasks prior to processing the data for destination sources like Google Analytics. They can rarely prevent from polluting your data. Proper implementation quality assurance should help you resolve most issues, but production environments tend to be the true final testing ground.
Post-processing data quality checks often takes the form of an tag management audit. Tag auditing solutions such as Tag Inpsector, OberservePoint and Hub-Scan have a well-proven reputation. Their methods, in general, involve the crawling of a website and analysing the Document Object Model’s content. This is done to determine which tags are loaded on a page, and more often than not, not the data in the tags themselves. Some solutions also scan, check and report on data passed to any known destination through network monitoring and compare this to a predefined set of data formatting rules. It can even go so far as to fix your tag management system in real-time.
However crucial these auditing solutions are, they remain dependent on crawling the correct pages in order to determing which tags are loaded on the page and what data is being passed on. The biggest challenge for these ‘outside-in’ tag audit solutions is the crawling itself. Crawling pages in checkout funnels, or pages hidden behind logins and cookie walls can be challenging leaving you with a siginificant gap in your audit’s results.
Improving your data quality, getting started
A second thing to consider is pricing. Migrating to tools like Segment or Harvest, to allow for data quality checks prior to processing, will require not only a significant investment but also a paradigm shift in terms of your company’s data collection strategy. Tagticians is a certified partner of both Segment and Harvest by Grain, we are happy to discuss to possibilities with you.
When taking the post-data-processing auditing approach, Hub-Scan and Observepoint are among the more expensive of the recommended solutions. However, combining tools like Tag Inspector and Verified Data is a fantastic way to get started in improving your data quality. When onboarding new clients at Tagticians we include the use of Tag Inspector and Verified Data by default as as a service.
Verified Data Audit Flow
So what is Verfied Data? Verified Data takes a more hybrid approach to performing data quality checks. Not only does Verfied Data spider a website and check tagging, it also checks data in Google Analytics. When accessing Google Analytics it performs an audit on the actual collected data. This data quality check is invaluable for detecting incorrect and possibly privacy breaching data in Google Analytics and potential sub-optimal configuration settings.
It is important to remember where these tools function in the data collection process. These tools cannot prevent the collection of bad data, they can only identify and notify data collection inadequacies. The most effective data quality solution should function at the source.
Data quality checks in tag management tools
Taking action after the fact is not the wrong approach. We encourage everyone to be active in improving data quality and consider using solutions like Verified Data. At Tagticians we believe that the future of data governance, including data quality and privacy law compliancy, is as an integrated feature in your data collection solution. It is essential with regards to your company’s Tag Management Maturity.
Segment and Harvest have identified this necessity and are in essence ahead of the curve. Verified Data can help you get started.
Segment and Harvest by Grain are two differenct types of solutions when compared to tag management tools like Google Tag Manager and Tealium IQ. Where the latter two allow for ad-hoc tag placement through its user-interface, the former requires thorough upfront planning and development. Both types have their pros and cons which we are happy to discuss with you in person.
Let’s be clear, there is nothing holding you back from building data quality checks into your preferred tag management solution. Google Tag Manager’s new Templates feature could help you improve your data quality, especially in your Custom HTML tags, but it requires a lot of time and effort from your side. That being said, we are confident in stating that built-in data quality solutions are the way forward.
Moving Forward
How do you move forward from here? Our answer is to just start spending time on improving your data quality. It starts with considering, in the first place, why you track what you track. Then determine if you want to perform checks in the pre- or post-processing phase. Don’t just implement a solution, agree on a goal and remain consistent.
Define, document, implement, check, monitor and contact us if you have any questions or require assistance in any way.
