BHLDN Google Tag Manager

The background

Back in 2019 at BHLDN, I worked with my team on a replatform project from an old Coldfusion based platform to a much more modern Ruby on Rails website, but that's not the important part. Previously, having worked with a platform vendor for our ecommerce platform, anything that needed to be added to the site (metadata, tagging/analytics, buttons, style updates) was added and maintained by the vendor's team. This was a gap we identified in the replatform that we wanted to take control of: the ability to implement vendor tagging and analytics when we wanted to rather than waiting for the agile cycle our vendor worked on.

We had some options at this point, leverage the enterprise setup of Tealium our parent company had (this would've been $$$) or work with our vendor to expose a data layer for our own setup. Obviously, we went with the second option or else this wouldn't even be a post.

The lesson

The technical setup is the easy part to explain (this was also 7 years ago and the website no longer exists). The harder part to convey is how much the change in cadence shifted my ability to experiment and learn. A few questions often came up when working on new campaigns: Is this worth a ticket? Will it even make the sprint? What if we get there and realize we asked for the wrong thing? So we'd hedge. We'd bundle requests together to justify the overhead.

Once GTM was wired up to a real data layer, that whole dynamic changed. A question would come up and instead of opening Jira, I'd just go answer it. A short while later the tag could be live. Sometimes I was wrong about what I'd find, but it didn't matter too much, because being wrong took an afternoon instead of eating a sprint. I could start actually poking at things and testing. I built real intuitions about the funnel because I could afford to be curious.

That's the part that's hard to put in a business case. It's not "we saved X hours on tagging requests." It's that we stopped flying blind and started trusting our own data, and that changed how we made decisions.

The takeaway

The biggest win was figuring out where how to separate the platform from the business needs. They provided the platform and stability of a refined platform. We got to handle our own marketing and analytics. Neither of us needed to be in each other's business.

Sometimes the cleanest piece of engineering you can do is just agree on where one system ends and another begins.