Why Semantic Layer is the next paradigm for BI

For the past ten years, we've had an analyst-centric workflow in BI, where analysts answer questions from the business and build content for others. This model absolutely works (I myself have trained some 40 analysts over the years), but it has some clear downsides. Give someone a dashboard and they immediately come back with more questions. As the demand to use data keeps increasing, playing the old game (essentially throwing more humans at the problem) is getting less and less appealing. That's why we got really excited about doing BI on the semantic layer, as it opens up a new way to think about BI.

Semantic BI

With a semantic layer, you can build a catalog of metrics and ship that to the organization. Using BI tools that are built natively for metrics, you can provide an experience that is so intuitive and flexible that your end users can now do their own BI. That means they can explore and analyze but also create content all on their own.

That's a change with some big implications - it means that we in the data team can take a step back and focus on the data model and the semantic layer and that all our end users can now do much more BI work themselves. You're essentially giving your end users a very flexible sandbox that has built-in consistency and - with the right tooling - is really easy to work with. We're seeing many of our stakeholders quickly fall in love with the flexibility and user experience we can now give them.

The semantic stack - dbt, Cube & Steep

Now is an exciting time to get into the semantic movement. dbt labs launched its new semantic layer at Coalesce in October, the standalone alternative Cube has developed into a mature platform, and there are now tools like Steep that are built natively for the semantic layer. The technologies in this new stack work well together, and skills in one translate well over to alternatives. As of last year, you can realistically go all-in and serve even a large, data-hungry organization using this new paradigm.

Join the semantic movement

• Are you onboard the analytics engineering train?
• Do you need to serve a lot of data-hungry users in an efficient way?
• Are you curious about the semantic way of doing BI?

All good reasons to get started with your semantic layer and rethink your BI for the next era.