As a software engineer, my overall experience using Looker has been honestly pretty solid, mostly because it caters heavily to a developer-first mindset. The absolute biggest game-changer for me is LookML. Unlike other BI tools where you are stuck in a clunky drag-and-drop UI, Looker lets me actually write code to define the semantic layer. Being able to use Git for version control, open pull requests, and keep the logic DRY with extends and explores makes it feel like a natural extension of our standard engineering workflows. That said, it is not entirely without its quirks. While LookML is incredibly powerful, the learning curve can be a bit steep for non-devs. This sometimes means data requests still bottleneck with the engineering team. Also, debugging complex Persistent Derived Tables or deciphering the exact SQL Looker generates under the hood can occasionally turn into a frustrating rabbit hole. You quickly learn that Looker is only as fast as your underlying data warehouse. Overall, though, I really like using it. It perfectly bridges the gap between raw SQL and non-technical business users. I honestly do not dread building business dashboards anymore, simply because I know the underlying data models are robust and maintainable.

It is a very solid tool, especially with the possibility of centralizing metadata and definitions with LookML, but the selection of visualizations available is very limited compared to similar tools from other vendors.