April 11, 2026
Measure wars go full SQL
Metrics SQL: A SQL-based semantic layer for humans and agents
Rill bets big on SQL for business numbers — fans cheer, skeptics yell “1998 called”
TLDR: Rill’s “Metrics SQL” lets companies define business numbers once and query them everywhere using plain SQL, even by AI tools. The crowd is split between “finally, sanity” and “we’ve seen this cube before,” with extra drama over letting chatbots near company KPIs — useful, but controversial
Rill just dropped a bold claim: their new “Metrics SQL” lets everyone — humans and AI bots — ask for business numbers in plain old SQL. Translation for the non-data crowd: define revenue or churn once, then use the same definition everywhere, so Finance and Product stop fighting over mismatched numbers. The blog name-drops a SQL legend and promises fewer mistakes, better security, and cheaper dashboards — and the internet promptly turned into a town hall.
On one side, SQL lifers are calling it common sense. “Finally, a single source of truth without learning another weird query language,” cheered one camp, tossing shade at past tools with custom languages. On the other, veterans of BI wars rolled their eyes: “Isn’t this just OLAP cubes with a fresh coat of paint?” Several joked that we’ve looped back to 1998, now with AI agents that “speak SELECT.” The AI angle split folks further: some love that large language models can query metrics directly; skeptics warn that letting chatbots touch company KPIs sounds like a compliance speedrun.
Memes flew: “LookML with vibes,” “SQL is the new English for robots,” and “my CEO’s favorite metric is ‘vibes per quarter’.” Meanwhile, a quieter middle asked for benchmarks and proof that this is faster and not just prettier. Buckle up — the metric wars are back, and they’re bilingual in SQL and sarcasm. Read more on Rill’s blog and Julian Hyde’s SQL world via Apache Calcite
Key Points
- •Rill built a SQL-based semantic layer (Metrics SQL) that treats metrics as first-class objects.
- •The approach aims to eliminate metric definition drift by using a single governed definition resolved via SQL.
- •The article critiques prior metric layers that introduced new query languages (e.g., GraphQL, MDX, JSON-over-REST).
- •Julian Hyde’s call for first-class SQL measures influenced Rill’s metrics-first design.
- •Rill defines metrics as aggregate measures in a dimensional context, enabling governance, security enforcement, and SQL-driven performance optimizations.