Melbourne Measure Camp 2020 was a great success, it was fantastic to see the event in full swing giving all attendees the opportunity to present and knowledge share in their field of expertise. Giving up a Saturday for some can be a big ask, however if you love data and analytics then it’s not really a hard decision at all - I mean honestly there are very few social situations that let me get into an excited conversation about the dataLayer and GTM.
Data Runs Deep is not only a continual sponsor for this event, we actively encourage our team to attend and most importantly get up and speak, sharing knowledge for us is part of what makes the company and culture great
It is always a privilege to present and personally something I love to have fun with. My slot was 3:15pm and my talk was on the dataLayer. For those unable to attend - or others who just want a refresh, here is the content in blog form. Enjoy! The content has been altered and I have provided discussion points and notes for the key slides
Cooking up a dataLayer
- This presentation will embark on combining my love of cooking and dataLayer design
- The goal is to provide key concepts of dataLayer design for scalability and re-use
- We will only focus on the dataLayer - not how its used in Google Tag Manager
- Before you consider data design you need to have clear business objectives
- Then you need to consider and evaluate what you need to achieve these objectives
- Let's consider the objective of a recipe - to bake a cake
- What comes to your mind when you look at the above imagery
- Are you detailed oriented or do you see the broad strokes?
- From broad to detail, what level is going to be adequate to achieve the objectives
- Data design is the art of labeling many different things in a coherent and logical structure
- It's about finding the right balance between detail vs being too generic
- In respect to a recipe (context) we can group these items under "ingredients"
- Each can have their own attributes, forming a deeper layer of abstraction, nested under the higher level.
- Does enhanced ecommerce product data sound a bit like this?
- With our data points defined we now need to consider the steps it takes to complete the end objective
- You need to define the expected journey and the key steps taken.
- These will most likely define the events and user actions to be tracked
- All journeys will have key steps, you need to decide how many of these are needed.
- You can add more steps in at a later stage, if your design is sound.
- Don't forget to include a start and an end stage, these will be constant to all journeys.
- Most applications become complex and are designed to optimise the users journey.
- Therefore when designing events and actions assume that they may or may not happen.
- They should be interchangeable , shareable or skipped completely
- Visualise the process in a table format
- Define a logical grouping for each component.
- Test the structure by changing out the values e.g. Banana cake to Apple cake
- If you have minimal changes to the structure you know you are on the right path
- A dataLayer sits between user actions, browser events and your tag management platform
- Your front end developers will need to implement it
- For the business - think of it as holder for data that the organisation wants to share
- It is a reliable and ideal way to pass data to tag management platforms
- event is the only required key, this describes the action that has occurred, and will provide context to the data sent along with it
- You can have any number of custom key & value pairs
- Visualise the recipe table under with the object key and value structure
- Converting the recipe table into a dataLayer push
- Generalise the keys
- You can now start to see the potential mapping to a Google Analytics Event, e.g. Category, Action (title), Label
- Using the updated structure and revisiting the table format, we can swap out the values for something more practical to digital analytics
- Example for a contact us form
- Contact us form dataLayer example
- Key dataLayer design summary
- Key dataLayer philosophy summary
- Think of the dataLayer as a beautiful cake you are presenting to your tag management. The platform then only needs to decide who gets a piece
- The alternative is to use the tag management to scrape the website, like raiding a pantry - grabbing what you can and mixing up a cake and hoping for the best. It may still be edible, but you're likely to feel a bit ill later on