In October 2019, Nike announced that it would replace its CEO of 13 years, Mark Parker, with John Donahoe. Rather than being from the sports world, Donahoe was a veteran tech executive with high-profile stints at eBay, Paypal, and the cloud computing company ServiceNow.
In hiring a tech leader for its CEO, Nike signaled a significant shift toward ecommerce, tech, and data-driven business. As Darren Rovell said, Nike wants to become “a technology company that happens to sell shoes and apparel.” An article on Harvard Business School’s site even asked, “Is Nike the next big tech giant?”
This move may seem strange for a sportswear company, but Nike is hardly the first non-tech company to put a tech leader at its helm.
In June 2021, Ferrari announced its new CEO Benedetto Vigna (a tech veteran with 26 years and 200 patents in the microchip industry) as part of an effort to move from luxury sports cars to newer innovations like electrical engines. Similarly, in March 2020, Siemens appointed Roland Busch (the former CTO who spearheaded Siemen’s digital push) as its new CEO. Known for creating equipment and hardware in the first industrial revolution, Siemens now sees cloud-based software as its future.
These companies’ embrace of tech reflects a new label for the 2020s — the Data Decade. Data analytics is becoming the core of every company’s business model as CEOs across industries are investing in next-level machine learning, artificial intelligence, automation, and hardware horsepower. As Forbes put it, now “every company is a data and analytics company”.
In my role as the founder of Atlan, I’ve been fortunate to spend a ton of time speaking to CEOs and CDOs about their data strategies. Over time, I’ve realized that across industries, there are four main ways that companies can use their data to create advantages for themselves and unlock greater growth.
1. Data as an operational advantage
In a story that has now become somewhat legendary, Shailendra Singh from Sequoia once recounted the story of Gojek’s first CEO, Nadiem Makarim, who was obsessed with the super app’s growth data.
Legend has it that board members of Gojek got not monthly, but daily updates with key metrics that the super app measured and movements from the past day. Makarim used these to track metrics on an almost hourly basis and form an intuitive sense of what could be breaking. This maniacal focus on using data to build an operational advantage fueled Gojek’s journey to becoming a decacorn and one of Fortune’s “50 companies that changed the world”.
Gojek’s story is one of using data as an operational advantage. This is the kind of data moat that every company in the world can build. It’s about understanding the key levers and metrics that drive your business, then using them to significantly improve operations.
An operational advantage means that CEOs and CDOs can answer key business questions in near real time, like “Are we meeting our daily KPIs?” or “Is there a drop in our conversion rate?”
A key aspect of using data to build an operational advantage involves making data available and understandable to those who are making daily decisions. This includes both technical and less technical users, everyone from data analysts and engineers to marketers and product managers.
Taking this a step further, major companies like Uber have found ways to democratize data — where people can not just find it, but can analyze and use it. Uber has made it possible for every user in the company to query data on an ad-hoc, dynamic basis. It’s estimated that over 400,000 queries are run daily on Uber’s infrastructure.
Other companies have even gone further by promoting digital literacy. For example, Airbnb created Data University and later Data U Intensive, data education for anyone at Airbnb that scales by role and team. After over 800 data courses and thousands of “butts in seats”, Airbnb has seen the percentage of daily SQL users spike from 7% to 62% and ad-hoc requests for the data team decrease by 50%.
2. Data as a strategic advantage
Every quarter, your company will make a few highly critical strategic decisions. If you’re a product company, this could be answering the question, “Which user segments should I be focusing on?” If you’re a hyperlocal logistics company, it could be, “Which cities should I be expanding into?”
These decisions can be made using varying levels of data maturity, but the more granular and real the data behind these decisions is, the more likely that the new strategy will help leapfrog the competition.
For example, in my past life, I had the chance to work with the top 3 oil companies and the Government of India to open 10,000 new LPG (liquified petroleum gas) centers.
Access to clean cooking fuel is a huge problem in India, so the Ministry of Petroleum and Natural Gas wanted to bring LPG stoves to 80 million women below the poverty line. Part of this initiative was opening 10,000 new LPG distribution centers — not just anywhere, but choosing the locations that would reach the most people in need and have the greatest impact.
We started with data from the 3 oil companies about the 17,000 existing LPG centers’ sales, locations, and customers. We then merged that with external data from over 600 sources about population, affluence, infrastructure, LPG penetration, etc. With this data, we could identify the perfect location for each new center, based on which villages didn’t have LPG centers nearby and which villages had enough market potential to support a new center.
3. Data driving a core product advantage
The third kind of advantage is when companies leverage data to drive a core product advantage.
The best example for this is Netflix. A decade ago, you could say that Netflix’s competitive advantage was its video software. Now, that’s definitely not the case. While video streaming services are common, Netflix has marched forward in part because of its data.
With years of user data, Netflix knows everything about each of our tastes, likes, and dislikes. Part of this comes from its not-so-basic user data — for example, tracking the completion rate, stop and start time, time of day, and viewing behavior (e.g., pause, fast forward, rewind, etc.) for everything we stream.
In addition, Netflix’s data comes from its micro-classification of different types of content — over 76,000 micro-genres even back in 2014. Netflix uses over 1,000 tag types that classify content by time period, plot, mood, etc. Instead of comedy or horror, think “emotional fight-the-system documentaries” or “period pieces about royalty based on real life”.
All of this user data from millions of users means that Netflix can personalize its recommendations to keep each person constantly hooked and binge-watching the latest shows. For example, the Netflix recommendation engine can predict that you’ll like show X because people who liked your favorite show Y also loved show X, supported by your viewing patterns of shows in related micro-genres.
This massive data moat — based on 14 years of streaming data for over 200 million customers — makes it hard for a new startup with better streaming services to overthrow Netflix.
Other examples of data acting as a fundamental business moat include Uber’s demand and supply matching algorithm and Gmail’s “smart compose” auto-completion feature — both powered by troves of proprietary data collected over time by these giants.
4. Data monetization: Data driving new opportunities
The last kind of business moat is turning your company’s data into a business opportunity by itself. While the age-old examples include ad networks like Facebook and Google, which leverage user data to deliver highly targeted ads, there are a ton of unique applications of data-powered businesses.
To take the Netflix example further, today Netflix is leveraging data about what people watch to create new TV shows and movies (called Netflix Originals) that are a perfect fit for its audience. This unique treasure trove of data that Netflix has access to means that it can give the best producers in Hollywood a run for their money.
“We always use our in-depth knowledge (aka analytics and data) about what our members love to watch to decide what’s available on Netflix….If you keep watching, we’ll keep adding more of what you love.”
– Jenny McCabe, Director of Global Media Relations at Netflix
For example, take House of Cards, the first TV series that Netflix produced. This was a massive investment for Netflix. It outbid major TV channels like HBO and AMC to get the rights to the U.S. version of House of Cards, with an estimated price tag of over $100 million for two seasons.
Why did Netflix take what seemed like a huge gamble at the time? The answer is simple. Netflix didn’t think it was a gamble, because its data showed that House of Cards would be a success.
Netflix knew that people who loved the original British House of Cards also loved movies with Kevin Spacey and David Fincher, and that one of its most popular movies (The Social Network) was also directed by David Fincher. A remake of House of Cards, directed by David Fincher with Kevin Spacey, seemed like a no-brainer.
Netflix also uses its data to make decisions like which of its shows it should cancel or restart, or which genres it should focus on.
For example, Netflix realized from its data that there was a gap in romantic movies, which Hollywood has been retreating from for years. Starting a few years ago, Netflix started producing a slew of romantic comedies — not just for Valentine’s Day, but year round. These include everything from teen flicks like Alex Strangelove and The Kissing Booth to more adult-oriented options like Set It Up and Like Father. The result was clear — Netflix’s romance movies exploded in the media, and over two-thirds of its subscribers watched a romance movie between March 2017 and March 2018.
So how can you start building a “data moat” or competitive advantage around data? In my next article, I’ll unpack a four-step prioritization framework that can help everyone — from a fledgling startup founder to a Fortune 500 CEO — to prioritize your data investments and create the maximum ROI.
Sign up for my Substack newsletter to receive the next article in your inbox!
This article was originally published in Towards Data Science.