Data quality and how to take responsibility

It happens, that I actually take my time to explain people what web analytics is and what it’s not. But what I should be explaining much more of
is the importance of data quality. The data you collect and make decisions from. The data that you will proudly use in presentations and have in
mind when you advance to your next job. But what is data quality? Most people, to be honest, don’t seem to care at all. They just want numbers.

I can understand why people don’t want to take their concerns on a higher level with data quality, not saying it’s okay, but I understand why
they don’t do it. It’s too complex and too time consuming to understand how data is collected and how it is processed. Many marketers don’t want
to take on a responsibility, that they think should be taken care of by a person in the IT department.

Well, time is up. We all need to adapt to a new reality.

With more power, comes more responsibility. In the next couple of months/years we will see marketers have more and more control over the data
collection in general, and that makes it even more important for them to understand “how” and “when” the data is collected. Have a look at Google
Analytics, Adobe Analytics, Tealium, Ensighten, Adform.. I could go on for ages, all these brands have tools that can give each and single
marketer the power to unleash a bad script that doesn’t fires off, because data collection is not something that is of any big concern. The
biggest concern to address with the upcoming and existing Tag management solutions is the governance. How can you ensure that a faulty script is
not being published? The solution is pretty simple. Make sure the scripts/tags are tested by your IT department before you release it to a
production environment. But how do you know if you are firing off the tag at the right moment? The truth is, you don’t know unless you know how
to debug your solution. As you can see, there is a gap here between the rules/tags being implemented and the business logic. In my opinion you
can’t be sure that the IT department will ask you the right business questions to ensure you that your data collection is correct, they will be
able to help you with deployment of the tag and ensure that it is firing, but you need to take control of it to ensure it’s firing at the correct moment, make sure the business logic applies as intended.

One of my best examples of this behavior unfortunately comes from a 3rd party, a media agency (no names mentioned), that solely is dedicated to
serve content on 3rd party sites and report back to the business based on the collected information. One of the things that nearly gave me a
heart attack was that they didn’t test the scripts after they were implemented on a clients website, they only looked in their own database and
concluded that the script was working. Now seen from a data collection point of view, that is insane. How can you even consult and make
conclusions based on data collection that you haven’t even tested for functionality and correct execution?

Unfortunately I see many companies do the same, all over the world, and the truth is that they have a limited understanding of the technology and
that’s where the pain point is. If you want to survive within the marketing field the next couple of years, do your self a favor, ask yourself
these two questions when you are working with data collection:

1. Do I know how the data is being collected?
if yes, good.
if no, then contact your IT department/IT Architect, someone needs to explain it to you.

Recommendation: There are thousands of different blog posts out there that explain how you can debug a solution.

2. When was the last time someone checked the tags/scripts?
if less than 1 month ago, then don’t worry.
if more than 2 months ago, you should definitely consider going through your most vital tags/scripts that have a impact on your KPI’s.

Recommendation: Plot in 2 hours in your calendar each month for only one purpose, to test scripts and tags.

And one last thing. Please don´t put the control of data quality solely in the hands of a third party. You need to be a part of the quality insurance process so you can make the final decision on whether the data collection is acceptable or not.