Dbt to acquire Transform to build out its semantic layer

Airbnb wrote a few great blog posts about why they built a standardized metrics layer - I think this one gives better technical context on the &quot;what&quot; and &quot;why&quot; than this announcement does:<p><a href="https:&#x2F;&#x2F;medium.com&#x2F;airbnb-engineering&#x2F;how-airbnb-achieved-metric-consistency-at-scale-f23cc53dea70" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;airbnb-engineering&#x2F;how-airbnb-achieved-me...</a><p>(As the article notes, Transform&#x27;s founders are Airbnb vets who worked on Minerva)

The interesting bit for me about Minerva (predecessor to Transform) was the engineering rigor that the pipeline allowed. It&#x27;s pretty easy to write a query that brings a DB to its knees and the source controlled and reviewed intermediary steps for producing metrics seemed like crucial features. If you then have a good catalog of what metrics exist you can let BI tools or business types with their own tools mix and match performant metrics to their hearts delight.

After 2 years of slow progress, dbt acquires Transform and their open-source MetricFlow to help build out the semantic layer.

Seems like a cool place to work, but I&#x27;m always a little weary of venture backed open source companies. When the board wants a liquidity event they&#x27;ll win over users every time- put more features behind a paywall or force everyone onto a SaaS platform.<p>Anyhoo, all pure speculation. While I&#x27;ve never used it, my friends have great things to say about DBT.

How does this compare to multidimensional OLAP cubes? Do they solve similar problems of different time windowed views of the data?

Can someone ELI5 why to use DBT instead of CREATE VIEW?