When I first started building out our GTM containers, data layer design felt outside what we needed. We were laser-focused on solving operational problems-tracking page views, basic events, and making sure the firehose of clicks didn't break things. The data layer was something I assumed the devs would sort out later on their end.
Fast forward two years and now most of our user interactions depend on structured data that never got built. We've created this awkward situation where the final attribution step of every conversion funnel still happens inside a black box. Looking back, I completely underestimated how deep the data layer ties into the entire product analytics experience.
Anyone else run into this? If yes, how did you retroactively fix it-GTM custom variables, pushing schema changes through a tag audit, or something else?
For context, the discussion that followed had replies like:
- We kept pushing data layer work down the roadmap because there were always more urgent tracking requests. Looking back, users (the analysts) cared about that gap way more than we expected.
- I think a lot of implementers go through this. You build around current event tracking and then realise the data layer is sitting right in the middle of everything.