Home Thinking Aloud Six Principles Of Data-Driven Transformation

Six Principles Of Data-Driven Transformation


by Nir Kaldero, author of “Data Science for Executives: Leveraging Machine Intelligence to Drive Business ROI

Even if you feel ready to turn your organization into a data- and model-driven enterprise, you may be unsure where to start. While a worthy ambition, a transformation of this magnitude is not a simple task. Many organizations are troubled and don’t know how to embark on the journey successfully. In this article, I’ll give you six principles you can use to guide your efforts.

These six principles will help you figure out where to begin and how to start the journey of transforming your organization, because becoming a data- and model-driven enterprise requires a plan as well as the right culture, people and process, data, and technology.

We’re now well into the Fourth Industrial Revolution. The First Industrial Revolution was about steam and railroads, the Second about electricity, and the Third brought about by the Internet. AI, the basis of the Fourth Industrial Revolution, will completely change the way business is done and companies are run in the next five to ten years, just as the Internet has done in the last ten. The transformation will be bigger than that any previous revolution has brought about.

The six principles described in this article are not a playbook or magic formula for your organization. They are based on and derived from working closely with many enterprises, from various industries, that have made the transformation successfully.

These are the six principles:

  1. Data strategy
  2. Data democratization, which is one of the biggest gaps organizations are currently experiencing
  3. A data- and model-driven culture
  4. “Speed to insight”: how to derive insights from your data as quickly as possible
  5. Data science value as a key performance indicator (KPI)
  6. Data governance, security, and privacy

Read on and I’ll explain how to make use of them.

Principle One: Data Strategy.

Data is the gold of the Fourth Industrial Revolution: your competitive advantage. Since your data contains hidden insights, you need to treat it as a strategic asset. Therefore, you should design a roadmap with a clear strategy on how to leverage your data and treat data as a strategic asset across and throughout your organization.

According to Ginni Rometty, CEO and Charmain of IBM, only “20 percent of the data is searchable.” The rest, 80 percent, is behind the firewall. This is your proprietary data and your competitive advantage. You already sit on a lot of hidden information about your customers, clients, and business that can help you transform your organization and take it to the next level if—and only if—you treat your data as a strategic asset informing all your business decisions.

Typically, when I talk about data strategy with business leaders, their immediate response is, “Hey, this means I’ll have to realign the entire organization. How would that work? How can I align all my 100,000 people with a single data strategy?”

However, setting data strategy is different from goal-setting. With goal-setting, we start at the top. Everything must orient to the goals top executives have set for the entire organization for the year. Data strategy, however, can be different for each sub-team and still contribute to the solution of your top business problems. These different strategies don’t need to involve a single set of constraints.

Principle Two: Data Democratization.

The second principle involves democratizing your data throughout the organization. This is important because everyone, from the barista to the CEO, makes business decisions on a daily basis. We know that data-driven decisions are better decisions, so why wouldn’t you choose to provide people with access to the data they need to make better decisions?

Let’s be practical, however. We live in a world of constraints and regulations. Not all organizations can completely democratize their data, particularly in industries such as banking, insurance, and healthcare. For privacy reasons, data leakage in these cases would be catastrophic. It would introduce direct business risk and liability.

The second constraint circles back to the concept that data is gold and a key to your enterprise’s competitive advantage. If data is such an important asset (and it is), you clearly don’t want to share all your data with the entire organization. If you did that, proprietary information might leak out and cost you your advantage.

So how can we democratize data intelligently? The answer is to figure out how to provide relevant data to relevant decision-makers so they can enhance their decision-making. Look at people’s roles, identify what decisions they make on a daily basis, and then provide them with the data that will support these decisions. Providing the right data to the right people will enhance their capacity to make the right decisions at the right time.

Principle Three: Building a Data-Driven Culture.

Principle three is about creating a data-science and analytics culture within your organization. Leaders must incentivize employees to cultivate the habit of looking at data whenever they make decisions, which I call “the point of action.” This is tightly linked to the corporate culture you build. I often suggest that executives get creative and set up competitions and rewards for employees who champion data.

A second component of this principle, and one of the biggest current gaps in industry, requires you to bring technical and non-technical teams closer together, working seamlessly to realize and operationalize machine intelligence. This is a key tool for the increase of ROI.

At this point, the gap between technical and non-technical teams remains significant. These teams don’t understand each other or know how to work together. This is a major problem that must be faced and overcome.

Nevertheless, there are early adopters, such as Google, which provide great examples of companies that have bridged the gap between its technical and non-technical teams. One of the remedies is educating both teams about each other’s roles and functions. The second is a smart, highly collaborative, embedded organizational work structure that requires the two teams to interact during the normal course of business. The third is creating a semi-technical role for a middleman (or middle-woman) between the two sides of the business.

Principle Four: Accelerating Speed to Insight.

The idea behind this principle is to democratize information and insight about your business throughout the organization. If you provide high-speed, dynamic insight to decision-makers, they will get into the habit of making data-driven decisions. The definition of a data-driven organization is an organization that cultivates a culture of looking at data to make all business decisions. To do that, it’s important to use your data to generate as much insight as possible.

One of the simplest and best ways to unleash insight throughout the organization is to use dynamic dashboard tools that provide insight into and beyond the data. Many organizations do not emphasize the importance and usefulness of such solutions. You already have ability to create dashboards that dynamically represent data and are simple to read and understand. Organizations need to move away from static summaries and reports. They are no longer dynamic enough to inform decision-making.

Principle Five: Data Science Value as a Key Performance Indicator (KPI).

The fifth principle of data-driven transformation is about taking action. You must measure the value and impact of data science and machine learning on your business and make this metric one of your key performance indicators (KPI). Always start with a small machine-intelligence initiative and investment, measure its success, demonstrate ROI, and then take on larger initiatives, while celebrating the wins and democratizing the knowledge throughout your organization.

In doing this, prioritize data-science investments with the highest potential ROI. A typical chief information or chief data officer at a Fortune 50 or Fortune 200 company receives between 2,000 and 2,500 requests a year for different data products. People within the organization think they should act upon all these, which is rarely feasible.

How should you prioritize? Look at an investment’s feasibility and impact. Feasibility refers to whether you have the data or not. Is the data clean and labeled? Do you have the talent, resources, and processes to get the project started? Impact refers to financial contribution. If you’re going to invest in this project, will it genuinely revolutionize your business over time? Will it add millions of dollars, or will it add $10,000?

Think about these two dimensions before you submit a request to your CIO for a project you think might be a good use case. Particularly when starting the journey, you don’t want everyone to submit hundreds of use cases. You want to grab one with high feasibility and impact that will be able to transform your organization quickly.

Start by piloting a project. If you see that a magnitude of change is reasonable, pour more money into it: invest more and hire more. Then operationalize it throughout the organization.

Principle Six: Data Governance.

This final principle is all about the environment in which your data sits. Your data assets must be secure and private. This is a priority, and all large corporations should have thoroughly established data governance, security, and privacy by this time. By my standards, however, many of the companies I work with are still quite far behind the curve. While the importance of safeguards should go without saying, it still needs to be said: many organizations haven’t yet instituted them.

This is a Pandora’s box that could kill an entire organization, as recent events have demonstrated.[2] Think about the repercussions. Humans, not machines, are responsible for data governance. Other people in the organization often treat those responsible for data governance like black sheep because they protect data. This is unacceptable. Security is a key principle in a data- and model-driven enterprise. It creates a healthy environment.

These Six Principles Will Move You toward Data-Driven Maturity.

Initially, applying these six principles may appear daunting. No doubt it will take you a while to start thinking about maximizing the use and protection of data in every decision you make. Nonetheless, it can be done. Indeed, I believe every large organization that wishes to adapt and thrive in the Fourth Industrial Revolution will need to begin using these six principles.

As an executive, positive transformations should start with you because they will trickle throughout the organization. You are a role model people look up to. If you start a journey, other people will join you. Before long, you will begin to see more people understanding and living by these principles. At this point, you will know your organization is on the way to data-driven maturity.

The journey won’t happen overnight. You always need to start with yourself: with your habits and your way of thinking, with hard work and deep intention (as I learned in my own yoga practice), the rest will follow.


*adapted from “Data Science for Executives


Nir Kaldero, author of “Data Science for Executives: Leveraging Machine Intelligence to Drive Business ROI“, is dedicated to bringing the benefits of data science and machine intelligence into business. As the head of data science at Galvanize, Inc., he has trained numerous C-Suite executives from Fortune 200 companies in how to transform their companies into data-driven organizations by applying the technology behind the “fourth industrial revolution.”


Please enter your comment!
Please enter your name here