Data As the Fourth Pillar
This article was previously published in Forbes
Throughout the history of commerce, from ancient-era merchants to 19th-century whaling ships to modern corporations, companies have used three main pillars to create strategic differentiation: people, processes and technology.
The past decade has been different. While the vast majority of companies have used or collected data simply in support of their technology pillar, an elite few have instead elevated data as a pillar of its own, finding success at historically unprecedented levels.
In his classic 1996 Harvard Business Review article “What Is Strategy?” Michael Porter makes an important distinction between “operational effectiveness” and “strategy.” To summarize, he defines operational effectiveness as performing the same business activities better, and strategy as performing different business activities than your competitors. In the context of Porter’s definitions, up until the past decade, data has been used for operational effectiveness but not for strategy.
Walmart was masterful at using the original three pillars — people, processes and technology — to dominate retail and reach the pinnacle of its industry. But then Amazon came along and elevated data into a top-level role as the fourth pillar.
Instead of simply competing against Walmart’s three pillars, Amazon massively invested in data management infrastructure to power numerous innovations such as its product recommendation engine and its ability to analyze consumers’ product research behavior. It then made its infrastructure available to others to create what is now the third largest ad-revenue business in the U.S., behind only Facebook and Google. Amazon’s investments in data management infrastructure caused them to suffer through nearly $3 billion of losses in its first five years as a public company before becoming profitable. Then it became really profitable.
By the end of 2015 — only 20 years after its founding — Amazon (with a market capitalization of roughly $294.5 billion) had nearly doubled Walmart’s market capitalization (about $167.5 billion). By the end of 2017, Amazon was on its way to doubling the combined market capitalizations of Walmart, Costco and Target.
You could imagine the boardrooms of those companies and others watching Amazon’s rise and commenting on it with the same mix of awe and bewilderment as golfing great Bobby Jones, who, after seeing a young Jack Nicklaus beat the best golfers in the world by nine strokes at the 1965 Masters, famously said, “He plays a game with which I am not familiar.”
It’s not that the Amazon executives were better operators than rivals’ executives. It’s that Amazon was truly playing a different game. In Porter’s vernacular, they were performing different business activities. They could do this because Amazon was working with four strategy pillars rather than three. It’s not unlike Einstein’s breakthrough in describing the physical world where all other physicists had failed, by adding the fourth dimension of time to the existing three spatial dimensions. Many existing beliefs about the physical world broke down, and entirely new opportunities opened up because of it. That’s the inflection point we are at in the world of business.
Microsoft, Apple, Amazon, Alphabet and Facebook all use data as the fourth pillar of competitive differentiation via sophisticated artificial intelligence/machine learning (AI/ML) deployments. As of this writing, they are the five largest companies in the U.S., representing approximately $7.2 trillion in combined market capitalization, or nearly 19% of the approximately $38 trillion entire market capitalization of all U.S. public companies. This is up from 0.5% in 1995, when total U.S. public company market capitalization was approximately $6 trillion and Apple and Microsoft’s combined market capitalization was approximately $30 billion, Alphabet and Facebook had not been started yet, and Amazon had just sold its first book. That is a 40-times increase in relative value in 25 years, and these companies are only getting more economically dominant at a rapid rate as they amass more data.
Now that the business world has seen this happen, the real question is: Why aren’t more companies using data as a fourth pillar? The answer is it is incredibly expensive and difficult to do this with existing data management tools, and only a small number of companies have the capital and knowledge to do it.
The foundational design of data management has remained fundamentally unchanged since the advent of relational databases and the three-tier application architecture, introduced nearly 40 years ago. Data was never considered to be a competitive advantage back then. Its purpose was to lie dormant in a database until called upon by application logic to support the automation of a manual process. Data management tools were designed to support that use case. To deploy the AI/ML systems needed to use data as the fourth pillar requires creating a “data-centric” architecture, MacGyver-ing the tools that were designed for the “application-centric” architecture that has been prevalent for 40 years.
Analysts generally estimate that the total cost of ownership (TCO) of making an application data-centric is about 10–15 times that of building and maintaining the core application itself. Research by Fluree (a company my firm invests in) found that the average application costs $175,000 for the database, app server and client software development, and the immediately identifiable “other” costs are about five times that. Layering in the cybersecurity measures adds at least another five to 10 times the cost of the core app stack in terms of products and labor.
Given that this is a cost, complexity and cyber-risk burden only a small number of companies can absorb, I believe the biggest investment opportunity of the next decade is in companies making the picks and shovels that reduce the TCO of building and managing data-centric systems, enabling their customers to use data as a fourth pillar of competitive differentiation. I previously wrote about how blockchain can address the cyber-risk burden, and we’ve already seen early versions of addressing the complexity aspects with Snowflake and Palantir, but we are still in the first pitch of the first inning in this game. Venture capitalists like me are actively searching for companies that address some or all these aspects as they have the potential to follow Snowflake and Palantir’s lead with multibillion IPOs over the next several years.