TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

From Data-Driven to Data-Centric: The Next Evolution in Business Strategy

In an age of data-driven strategic decision-making, some companies are sensing a shift: to be successful in the future, they will need to evolve to be data-centric, not just data-driven.

• By Troy Hiltbrand
• May 9, 2024

In recent history, companies both domestic and international have focused on becoming data-driven. They have mastered consolidating, managing, and leveraging data to inform their strategic decision-making. Many companies use a combination of descriptive, diagnostic, predictive, and prescriptive analytics to enable decision-makers to successfully guide their operational direction in an increasingly digital world.

For Further Reading: Building a Modern Data Management Strategy with Dr. Joshua Stephens How to Monetize Your Data Assets Data Platforms for AI and ML with Richard Winter

Now being data-driven is no longer sufficient. Companies preparing for the future have begun to evolve themselves to use a data-centric strategic approach to run their businesses. The question is what does that mean, and how can you be on the leading edge of sensing the shift and make the shift to data-centricity.

There are three key pillars of a data-centric strategy: data as a product, data as an asset, and data as a platform. 

Data as a Product

Once a mature data infrastructure is in place to gather, curate, and expose critical information internally for digital decision-makers, the next step is to evaluate whether any of that data can be monetized and sold externally. Data monetization has often been the realm of a small subset of data brokers. These data brokers have focused on commonly demanded data points that are not easily accessible to businesses but that they need to fill gaps in their data analytics efforts. Examples of these include market data (such as current stock prices and exchange rates), demographic data (such as customer preferences, associations, and income level), and scientific data (such as weather forecasts). These data brokers have built their business around data as a product. They know how to package it and offer it up to their clients for consumption securely and reliably.

As we move toward a data-centric business strategy, companies beyond data brokers are starting to ask themselves if they have what is needed to offer their data as a product.

As a business, are there data points that you have that would be valuable to others? Do you have a way to expose these data points without increasing operational risk? Is there a market willing to pay for your information? These are the types of questions that data-centric organizations are starting to ask themselves and discussions that are starting to happen at senior levels. Data monetization will play a bigger role in establishing an organizational strategy and will put an increased load on data teams to ensure data quality and secure methods to access the information.

Data as an Asset

With traditional generally accepted accounting practices (GAAP), accounting for data on the company’s books has not been easy. This has led to differences between the book value of an organization and the market value of an organization, but this difference demonstrates that data has value.

It is hard to quantify how much your data is worth and what the company’s efficiency is in translating that data into hard value for the organization. Companies are having to re-evaluate this challenge and get creative about how their data can be classified as an asset. As part of this discussion, there are varying levels of asset accountability for the data. When data is in its raw format, there is potential for it to ultimately generate value, but its worth is highly nebulous at this early stage.

As that raw data passes through multiple stages of refinement, consolidation, transformation, and augmentation, the value of that asset increases. This is similar to the processes that happen between raw materials in a factory and finished goods. Just as a manufacturer does with physical raw materials, data-centric organizations should be able to assign value to the data at each stage of refinement.

During the data monetization process, companies will need to determine if the value associated with the raw data as it passes through its multiple stages toward a finished goods state is merely a contributor to market value or if there is a potential to price and sell that data to others. Could your raw data or partially processed data have value to a partner organization? How do you account for that value and how can that be represented alongside your physical assets on your balance sheet? It is these types of questions that will be a central focus in data-centric organizations.

Data as a Platform 

Finally, we are seeing a new paradigm with the advent of generative-AI-based companies -- namely, data as a platform. Enterprises are not necessarily selling their data, but are selling the capabilities associated with models, such as large language models, that were built, trained, and refined on their data. The data enables the platform, and these companies are selling access to the capabilities associated with these platforms. When users of ChatGPT, Anthropic Claude, or Google Gemini create and run a prompt, they are not getting access to the data itself. Rather, they are getting output from models built on billions of data points.

These large language models (LLMs) are general and can perform well on a wide variety of prompts. With the concept of transfer learning, companies are using these base LLMs to create industry-specific models that perform very well within a single domain. Data-centric organizations are looking to see if there are opportunities for them to build out platforms that leverage their proprietary data to extend the base and provide business enablement processing to their clients. They may consider wrapping the base models with context-specific insights and data points or regulatory and security protections, or creating customized interfaces atop their data-powered platforms as a competitive differentiator. 

A Final Word

As companies start to shift from being data-driven to data-centric, the questions they ask about their data shift. They will transform from having discussions about enablement through data on their strategic decision-making processes to discussions about how they will value and surface the data they have in new and innovative ways as competitive differentiators.

Data-centric organizations realize that data is not just the asset that drives their business in the future -- it becomes the central component of their business in the future. Although this will not happen overnight for all companies, it is an important discussion to start having at the executive level to ensure that your organization has relevance in an increasingly data-centric world.