Data-Driven Marketing Starts With the Data Layer

The term “digital marketing “ should describe the marketing methods as much as it describes the marketing medium.  

Too often, marketers are satisfied with their “digital marketing” efforts when they send a tweet, “blast” an email or run a search ad because their message is transmitted over a digital medium.  Unfortunately, this misunderstanding so vastly underestimates the potential power of digital marketing; well…I just had to write a blog post about it.

If you take just one idea away from this blog post take this: The  access to data that the digital medium affords, when interpreted correctly and acted upon confidently, infinitely strengthens the performance of the digital marketing medium.

If you take two ideas away from this, take this too: The data layer is the very first place to start to capture and leverage your data. It provides the capability to merge data at its source. This is the foundation for data-driven action in real-time.

I wrote this post to illustrate the value of the data layer in the context of today’s digital marketing stack. I explain the present state of the digital marketing medium and how and why the data layer has a role in it. If you’re done reading, please tell your friends. The share buttons are above. Otherwise please read and leave a comment below.

I deal with data

I have data skills!

What is the data layer?

It is essentially a menu of available data, from which data consumers (defined and explained later) can order from within your web app or page.  Technically speaking, the data layer is a JavaScript data structure that receives and holds data about the user, the user’s interactions, the app, the view or page, and the context of the user interaction within an app or page.  The data layer can then be accessed by a tag management system or application for data collection or to trigger marketing actions.

The tag management system sits on top of the data layer and is responsible for routing data to data consumers. This could be as simple as a few lines of home-baked of Javascript or as intricate as Google Tag Manager. Ultimately, a tag management system is the logic that decides where the data is routed and to some extent how the data in the data layer will be used.

Why does the data layer exist?

To understand the role that the data layer plays in digital marketing, it helps to understand two currently sexy terms: Web App and the Digital Marketing Stack. I will explain those later but let me pull a little Chuck Palahniuk on you and set you up with these three eerily similar but presently unrelated premises:

The data layer exists because:

  • Web apps have multiple players providing and consuming data within the app.
  • Similarly to the concept of data independence, data access should be flexible and independent from the external user-facing layer of an application.
  • Data access and acquisition should not be barriers to data-driven action.

And now a story…

How the data layer came to be

(and A Brief History of the Web Apps.)

In the beginning there was HTML. It was and is the commonly used document structure for web documents.  In a very basic sense, the HTML language codes how text is hierarchically structured in a document. So as you remember in 1997, we could request an HTML document from a server far away and we could view it in our browser as a webpage, and things were great.

Sometimes, people would add some style to these documents and, depending on the browser you were using, things may or may not have been so great. The problem with adding style to an HTML document was that writing the style rules to the document itself meant that any time there was a need to change the style of a document, it meant changing the document itself. This was very messy and inefficient.

Then came CSS, a separate additional layer of style rules that your browser would apply to an HTML document.  Colors, sizes and blinking distraction exploded on the page in a barf of glory. People in the Wild Wild West of the web now had a simple method of keeping their theme of yellow Courier font over the top of a picture and John Wayne across all the pages of their site by using the same style sheet across their site. It wasn’t always tasteful and it certainly wasn’t interactive but it was consistent. Web developers hailed the separation of style and content layers.

Hello World! Here comes JavaScript!…and ten more pop-ups about things you wouldn’t bring up at a dinner party. JavaScript could both blow your mind and ruin your browsing experiences depending on your browser. Like document style, it was often a jagged mess until people realized the virtue in separating the JavaScript code (the behavior of the page), from the content and style of the document. Recognize the theme?

JavaScript is what is responsible for the prolific < booming voice > WEB APP </ booming voice >.  It is behind all of the shifting, scaling, constantly updating, interactivity that we now know and love. Possibly, the most powerful aspect of what JavaScript offers is the ability to get more data from a server without requesting a whole new web page. This is called AJAX and is what makes Facebook login for your favorite ecommerce site possible. It is also what makes web analytics products like Google Analytics possible.

Today, interactive Web Apps have become the norm. From Facebook, to Gmail, to EBay, to your website, web apps do more than just serve dynamic content from a server. With the help of AJAX, they can host an entire ecosystem of plugins, pixels, scripts, snippets, iframes and APIs that communicate with the other servers of the web! There are data providers (players that add data to the page like an embedded YouTube video, Qualaroo survey or the app) and there are data consumers (players that take data on the page for their own use such as analytics platforms like Google Analytics, MixPanel and SiteCatalyst, marketing tags like Adwords, Facebook or Adroll, email marketing platforms like Campaign Monitor or Mailchimp and testing and optimization platforms like Optimizely using all this data to trigger their own events using logic defined by each specific data consumer.)

“There ought to be a layer!” some smart person shouted after this mix of independent data and logic got pretty messy.  Again, a new web feature arrived and a new layer arose.Each layer of an application separated the business of each feature from the business of other features. By separating all the layers, it became possible to manipulate and modify each layer without affecting other layers or the application as a whole. The web is stable. Occam put his Razor away.

Enter the data layer. There is data in the application, there is data in the users’ interactions, and there is data coming from outside sources. The data layer exists to separate all this data from the business logic of the application. This way the application is only responsible for serving the user, and all other data consumers have a central point of access for the data they use.

How the data layer works (by example)

Let’s say that your brand just produced a video to promote a new product.

Your digital marketing stack looks like this: Your brand’s website uses Google Analytics to track user interactions and session source from several paid and earned channels that you cultivate. You have remarketing tags for Facebook, YouTube and Adwords and you use Optimizely for testing and dynamic content. Finally, you use Google Tag Manager to manage and trigger your marketing tags.

Because you’re smart, you recognize that just posting this video on YouTube is not going to have any impact on sales. (The right people have to see it!) You also recognize, based on your last video campaign, that users who saw the video were 15% more likely to convert, making it very important for users to view the video.

So you feature your video on the homepage.  As users come in, you use JavaScript to check their referrer and place that information in the data layer.  Based on the referrer you can infer their likelihood to convert; we know that a user who comes in via a certain branded paid search campaign is very likely to convert. We tag users who visit from this campaign with a value of “high” and place that data in the data layer. You could capture this in Google Analytics in a “User Value” custom dimension.

As users interact with the home page, some watch your new video and are tagged as “video viewer” (again, this could be captured as a Custom Dimension) and some navigate past the home page without viewing the video.  These users are not tagged. Because users of this segment have proven to be worth such a high value based on their referral source, you have set up Google Tag Manager to tag this specific audience with both Facebook and YouTube video remarketing tags. This user must see this video!

A couple days later, the user returns via a Facebook video post tagged with UTM parameters that are again read by JavaScript. This time, because the user came from the video post, we use Google Tag Manager to tag this user as a “video viewer.” When this happens, another event happens: this user is added to an Optimizely audience. Optimizely then cookies this user so that when they navigate back to the home page, instead of seeing the video front and center, they see a persuasive call to action above the video that shortcuts them toward the ultimate conversion page. Finally, a transaction is registered in Google Analytics along with a few events and a branded search campaign and Facebook video campaign source for attribution.

In this case, there were four players involved, but the most important thing to recognize is that without the data layer to unify all the data that each player was consuming, each player would have continued to act independently. The data layer provided the capability to merge data at its source, making it possible for all these players to interact with each other in real-time. The value created by each part of the digital marketing stacks working together is greater than the sum value of their parts.



Value is in the aggregate. Be a data gangster.

So much data! So much opportunity!

There is data everywhere. As long as there is a common thread between a single user and your website/application, there will be data that describes that relationship. Managing this data, finding the value in it and finding the value that it provides to other data role players is the key to maximizing digital marketing success.

To learn more about the data layer in action, see: Enhanced Ecommerce Using the Data Layer.

Leave a Reply

Your email address will not be published. Required fields are marked *