The buzz around metrics and semantic layers may have died down but the core ideas and work carries on.
Almost 2 years ago, Benn Stancil wrote an article that may have single handedly stoked existing sparks into a raging fire: that we need a metrics layer to define, organize and distribute metrics to downstream consumers and tools. The image shown is taken from his iconic blog post.
The original formulation emphasized governance, and was envisioned as an infrastructure, middleware layer.
While the "one source of truth" formulation is intuitive, realizing its value was never going to be quick and easy, as it requires pulling together an ecosystem to embrace standards. If standards are not adopted, the value realization slows down, and vice versa.
But while the overt excitement around this idea has dissipated, the work continues. DBT continues to invest in the core "standards" idea, while other tools, both new and old, have embraced and internalized these ideas into their own back-ends.
Personally, I am still excited about what standardizing metrics enables, regardless of how it is practically implemented.
Metrics are how data investments finally come to life within organizations. They are how business strategy and goals are determined. Metrics and segments are the lifeblood of day-to-day analytics and operational workflows.
As businesses organize and standardize metrics, governance and reliability may not even be where the value realization begins and ends.