The role analytics plays across organizations has evolved—and it’s the modern data stack that’s responsible for entering analytics into a new era.

I recently sat through a presentation in which a key section was highlighted “Data Visualization is Dead”. The notion that data visualization, business intelligence, or the dashboard itself is a dying medium has increasingly taken prominence. It seems odd to me that the vehicle that allows us to interrupt raw, disorganized data points and present them in a way that enables better decision making is now considered obsolete by some—especially at a time when understanding the story that data tells us has become essential to success.

But as with anything in technology, the new becomes old, and even newer ideas take shape. It wasn’t long ago that the in-vogue industry notion was taking the power of development out of the hands of “IT” and into the hands of business users. Data democratization was the North Star, and speed was the impetus of this evolution. Business users needed their reporting quickly, and the time and investment in a well-formed data warehouse could sometimes be perceived as a bottle neck. Business intelligence tools commonly stood in as the ingestion, replication, transformation, data warehousing, and data visualization layers. The time to value driven by BI tools serving as semantic, transformation, and visualization layers across the enterprise was of paramount importance. What was lost in this approach was an eye on scalability, and an easily traceable path to what constituted the single source of truth. This need gave way to the modern data stack we know today and the evolution of the dashboard.

So, to be clear, I don’t think data visualization or analytics is a dead—or even close to a dying—industry. But I do think the role in which analytics plays across organizations has evolved, with the adoption of the modern data stack serving as the battering ram. To understand how this thought has gained increasing visibility, is to understand how the industry has shifted in a relatively short period of time.

The Evolving Role of BI and The Dawn of The Modern Data Stack

Understanding the historical trends of the past few years is critical in framing the “is the dashboard dead?” question.

Business Intelligence 1.0, as it’s commonly defined, occurred in the 90s as the cost of data warehouses declined and concepts such as ETL and OLAP gained popularity. Tools such as MicroStrategy and Cognos gained prominence and the concept of BI reporting became commonplace within organizations. In this model, performance of the BI tool was tethered to that of the data warehouse. Business Intelligence 2.0 ushered in an era in which speed was of utmost importance. Long refresh intervals on the part of the data warehouse, and the general unease around latency, led to developments within BI tools that offloaded the ETL processes to the BI tool itself. This remained a popular method until the advent of cloud data warehousing technologies. The next major evolution would be the advent of cloud computing.

With the ubiquity of cloud data warehousing technologies, it’s hard to believe there was a point in time where data warehousing was the domain of traditional, on-premise deployments. What drove this shift was the realized savings and flexibility that cloud vendors offered. As more organizations reevaluated their cloud strategy, the role of BI tools was increasingly in the spotlight. Years of data logic being built into universes—or proprietary semantic layers—was increasingly being called out as an untenable path forward.

Data visualization, analytics, and dashboards are not dead, but the way in which they’re deployed and consumed has fundamentally changed. This has been felt in two ways:

1.) Business Intelligence is a Visualization Layer: 

While this sounds simplistic, it’s a subtle shift. Our client base is gravitating toward the approach in which ingestion, replication, and transformation logic are the domain of the cloud DWH. The data logic and transformation processes within a BI tool are kept to a minimum, and BI is truly just a data visualization layer in this scenario. This approach fits to the concept of Headless BI or “Bring your Own Analytics Tool” in which data visualization tools can easily be interchanged as the semantic layer is decoupled from the tool itself. The rationale behind this is flexibility in approach. If layers of transformation logic are built into a reporting tool, it makes it fundamentally more difficult to unwind that logic or move off that tool in the event a change is needed. The prevailing logic is if core data logic, and the optimized reporting layer, is kept entirely separate then one or multiple BI tools can be deployed easily as needed.

2.) A Gravitation Away from Canned Reporting:

Standard dashboards and reports fit to the concept of guided analytics. Users log into a report, view their key KPIs and visualizations, take the information needed, and move onto another process to take action. The concept of embedded analytics—or placing operational processes within reporting to take immediate action—has increasingly become popular. This is a shift away from traditional canned reporting, in which users can view their insights and take immediate action (i.e., update records, initiate a sales process within the tool itself) as opposed to moving out of the BI tool to initiate action.

The dashboard’s role in the modern data stack has become increasingly segmented in the approach to development. Larger enterprise customers, with the time, budget, data volume, and talent pool needed can feasibly adopt cloud-based technologies to fit to the Headless BI concept.

However, there is a segment of the market in which those same constraints are real concerns, and in those cases data visualization tools still serve as both the semantic modeling layer and visualization layer. Framing the conversation in a binary, or one-size-fits-all context is a lazy argument—and at worst, disingenuous. To not acknowledge, or plan for, the evolving role of BI is to be left behind. But to make a blanket assumption that data visualization is dead is to ignore how quickly industry trends shift, and that the “modern data stack” can have vastly different definitions for organizations.

Data Visualization is Here to Stay: It Just Looks Different

As long as data exists in an optimized reporting layer, or data warehouse, there needs to be a mechanism to view insights or initiate action. As long as that need exists, then data visualization will never be dead. If I have any advice, it’s to keep your eyes and ears open but to not let larger macroeconomic forces or industry trends be the end all determinant in your individual approach. Find the approach that fits to your organizational needs and ensure that approach is well-architected.

So no, the dashboard definitely isn’t dead. It’s simply evolved in how it’s consumed, how it’s deployed, and how it fits within a modern data stack. And in another ten years, when the industry winds inevitably shift, it will be something entirely different.

Kevin Lobo Kevin is our Vice President of Analytics and is based out of our Chicago office. He leads our Analytics service line, overseeing its strategic direction and messaging, while ensuring delivery of high impact analytics solutions for our clients. Outside of work, Kevin enjoys spending time with his wife and daughter, as well as running and live music.
Subscribe to

The 8 Update

Sign up to receive our monthly newsletter, and get the latest insights, tips, advice.

Thank You!