Recently, I read an E-book authored by Jason Call from ObservePoint on the topic of Data Layer and found it mighty useful as an introduction for those who still have no idea what a Data Layer is.
For those who are not familiar with ObservePoint, it is a software to help companies ensure digital analytics data is accurate and actionable. From the get-go, it might not sound very impressive, but that’s only because most companies have not understood the importance of having clean, accurate and actionable data, and the danger of working with bad data.
Fancy making awfully bad business decisions that will send the company down the wrong path? Then don’t bother with ObservePoint. Otherwise, I personally feel that it is a great addition to any Data Governance team.
Oh, just so you know, I have no affiliation with ObservePoint and neither do they know I am writing this post. That should clear things up.
And also, this article contains a great deal of materials from Jason Call’s E-book and a dash of my own expertise to provide a fuller picture of what Data Layer can truly bring to us.
Vendor-specific data collection
The E-book started off with this picture showing thousands of marketing technologies out there. Companies might not be using all of them, but I am sure most companies are using at least 10 Martech to stay operationally relevant.
Martech usually gathers data through the usage of tags. And to get those tags going, the data must come from somewhere.
Here we can see a tag in action, where we have to put the values into those weirdly-named variables (for those savvy enough, that’s an Adobe Analytics tag!).
Imagine doing that at least 10x over on your website.
Let’s not even talk about website speed, the idea of bloating your website with such unglam code is horrifying. What if I decide to change the page name from “Shop>Holiday” to “Shop>Summer Vacation”? Yeap, we will have to change that as many times as the number of tags lurking on our website. And that is provided you can remember where each of them is. Good luck with that.
And seriously.. it takes a memory-master to remember exactly what each variable represents. Names like s.eVar55 and s.prop1 are not acceptable at all. Even the geekiest of developers have the courtesy of giving each variable a meaningful name.
Data Layer to the rescue
Data Evangelist Justin Cutroni explained the data layer as follows:
The idea is akin to a dictionary where we store all the different words made available to us. So instead of each and every one of us coming up with our own dictionary, we now have only one that everyone understands. How great is that?!
With data layer, all vendors can now pick and choose whatever data they need to make their technologies work. If there is something that is missing, just add it to the data layer. You never know who else can benefit from that piece of data too!
Now with data layer, we have:
- Much more intuitive naming convention
- Data container that is vendor-agnostic
- Values managed centrally, making changes much easier
- No additional page bloat
- Increased data consistency across all Martech used
- Happier web developers
Other than curing cancer, data layer has everything someone is looking for to manage data in a clean and efficient manner.
Jason Call brilliantly talked about why data layer should replace DOM-scraping.
DOM-scraping is like borrowing your friend’s Netflix account because you are too cheap to get one. Most of the time, you have full access to uninterrupted binge-watching capability. For some reason, your friend decided to change the password. Your access is now broken, leaving you crying like a baby and blaming your friend for his attempt to secure his account with a stronger password.
DOM-scraping is just that.
Web developers build and enhance websites for a living, and we cannot expect them not to modify a website just because we selfishly decided to obtain values by scraping the webpages. Instead of limiting what web developers can do, we empower them by providing them with a data layer where they can store the relevant values for our use. Win-win!
Implementation and best practices
Almost a year ago, I have written a piece on data layer myself.
To better understand how to implement a proper data layer for your company and what are the best practices to follow, be sure to read it!
A big thank you to Jason Call and ObservePoint for continuously seeking ways to improve data governance for web analytics. Web analytics is still pretty much a nascent technology in my eyes, where most companies know about it, but few were able to reap true benefits.
Let’s not be a victim of “garbage in, garbage out” and start taking data governance seriously today.