4 Key Takeaways from the Semantic Layer Summit

We recently hosted a virtual Semantic Layer Summit with our friends at Snowplow, Databricks, DAS42, and InterSystems. During the event, we gathered data leaders and top industry technologists to discuss the importance and impact of using a semantic layer for data and analytics.

In my keynote address about “The Semantics of a Semantic Layer,” I kicked off the virtual event by highlighting a few keywords in the definition of a semantic layer that speak to its ability to promote self-service and provide control, consistency, and agility to any business. Throughout the day, our experts built upon those concepts — exploring how a semantic layer enables, evolves, and scales key initiatives to improve the way we integrate data into business decisions.

Below, we explore some of the primary themes that were discussed during the event. Interested in learning more? Check out all of the session recordings here.

1. The Semantic Layer is Critical to Scaling Enterprise AI & BI Initiatives

In my keynote session, I likened the semantic layer to a glue that bridges AI and BI. A semantic layer enables businesses to combine descriptive, diagnostic, prescriptive, and predictive analysis on one platform and break down silos between data science and business analysis. It’s a powerful collaboration point that scales AI and BI initiatives for any enterprise, helping to maximize value, increase ROI, and drive more business impact.

That sentiment was echoed in more sessions throughout the day. In his “Bridging AI & BI” session, my colleague Daniel Gray showcased how a semantic layer can help BI teams benefit from each other’s work — creating a flywheel that contributes to compounding knowledge for the whole organization, ultimately increasing time to insights. Senior Data Engineer at Trumid Colin Reid also touched on how the semantic layer will continue to increase the pace of innovation as ML and AI move into more parts of the business over time. Specifically, he notes that as business leaders grapple with explainability, a semantic layer will enable better collaboration and push enterprises to continually move faster.

For more on this topic, check out these sessions, now available on-demand:

The Semantics of a Semantic Layer
Actionable Insights for Everyone
Market Take on the Semantic Layer
Unleashing the Power of a Semantic Layer

2. A Semantic Layer is the Key Enabler of Data Mesh

During our panel about “The Evolution of Data Architecture & The Semantic Layer,” Benn Stancil acknowledged some confusion behind the concept and definition of data mesh. At AtScale, we’ve defined data mesh as an organizational structure for data that is business-centric and an enabler for self-service analytics.

Personally, I like to refer to it as a “hub-and-spoke” approach instead. We move too fast these days to leave data in the hands of a centralized source, but full decentralization is not a great approach either — because too much can be lost in translation between teams and tools. A “hub-and-spoke” approach combines the best of both worlds and delivers on the promise of data mesh. With this approach, a central data team owns the data platform, tooling, and process standards while business-embedded data stewards own data models for their business domains.

The catch? This isn’t possible without a semantic layer. A semantic layer operationalizes this emerging approach/methodology with a domain-oriented, self-service design.

For more on this topic, check out these sessions, now available on-demand:

Designing and Executing a Scalable Enterprise Data Strategy
Data Engineering & Architecture
Data Platform & Insights Tech
Data Fabric + Semantic Layer: Gaining Value From All of Your Data
Enterprise Analytics & Data Strategies

3. The Semantic Layer is Critical Part of the Modern Data Stack

Throughout the Semantic Layer Summit, many speakers touched on the evolution of data architecture. In our panel about “The Evolution of Data Architecture & The Semantic Layer,” Benn Stancil provided an overview of today’s data tooling landscape. He made an important note that the industry’s efforts to pull apart pieces of BI with new tools had an unfortunate consequence — we left issues related to semantics to the side.

Kirk Borne elaborated on that point and reiterated the impact of silos. Silos have led most enterprises to struggle with data democratization. Companies keep data in different data marts and data warehouses — and use multiple BI tools to consume and prepare reports. It’s a governance nightmare that leads to conflicting analytics outputs and stale data reports.

With a semantic layer, he says, we can right the wrongs that the industry inadvertently created. A semantic layer helps bring together everything that had been previously pulled apart — it’s a crucial piece in bringing insights together and making insights more actionable.

For more on this topic, check out these sessions, now available on-demand:

The Evolution of Data Architecture & The Semantic Layer
The Semantic Lakehouse
The Data Cloud

4. The Semantic Layer is Key to Data Literacy

Another common theme throughout the Semantic Layer Summit was the importance of data literacy. Jordan Morrow, also known as the “Godfather of Data Literacy,” joined our main stage session about “Scaling Data Literacy” and defined the concept as the ability to “read, work with, analyze, and communicate with data.”

Building on that concept, fellow panelist Chad Sanderson walked through a semantic graph project, visualizing the way a semantic layer creates a consistent language for business data. As his graph shows, consistent definitions are the key to helping teams throughout an enterprise understand and communicate with each other. By mapping data assets to the context within the real world of a business, more people understand that data and can tell a story about the health of the business.

For more on this topic, check out these sessions, now available on-demand: