The biggest opportunity in retail now isn’t something sold in stores to consumers.


As the world reopens, retailers and investors are asking: What’s the best opportunity in retail right now? A traditional answer would be something like: “sweaters” or “new cars.” But traditional answers aren’t the answer right now. While we see consumers willing to spend more than in the past on certain products like outerwear, footwear and underwear, that’s not it either. The biggest opportunity in retail now isn’t in a product at all, it’s in improving how things are sold. The biggest opportunity in retail right now is: retail technology.

But… Which Technology?

Imagine all the times you picked up a product, examined it and put it back, or you looked at a product online in detail and closed the tab without buying. If a retailer could aggregate your activity with millions, or even billions, of other consumers, they would have a much better idea of how to improve performance and your satisfaction. Technology like that requires intelligence and analysis that is so massive it is way beyond human capability. That kind of capability, the ability to make human-like decisions, is what’s now called artificial intelligence and it is now being used by hundreds of companies to facilitate a better customer journey in stores and online. Artificial intelligence is possible because of another technology, machine learning. Together, they are the underpinning of massive changes and improvement in consumer experiences to come.

What is machine learning and why is it important? Look for a moment at the objects around you right now, like the chair you’re sitting on. You may be seeing it at a different angle from others you’ve seen but your experience allows you to understand what it is. Machine learning learns like you, from experience, but more crudely. I am told that it takes about 100,000 images of an object for a computer to machine learn and identify an object like a chair consistently. Having access to 100,000 or more images of one type of object, and doing it for millions, or billions, of objects, is a long and expensive process. The sooner a computer can access and analyze large amounts (coders call it a “lake”) of such images and information, the faster it can integrate that image and “learn” what is going on around it, recognizing, for example, the difference between a chair and a table, between a living room and an office or between an apple and a steak on a supermarket checkout belt. The software that can access all those images can use machine learning to improve its capabilities autonomously. The software that has access to the biggest data lakes will make better decisions over time because it will do what we humans do: learn from mistakes. The software makes itself better when it learns more, just the way humans can.

Software with access to the biggest data lakes pulls ahead of competition. In artificial intelligence, the most successful companies aren’t necessarily the ones that have been created by better coders, they’re the ones that have gained access to bigger data lakes that allow them to self-improve faster. As long as the programs keep accessing more and more information to self-improve with, other programs can’t catch up. That aspect of artificial intelligence, the way that software can self-improve as it analyzes information, is new.

The software developers who can assemble and access the biggest data lakes are going to win because their software will pull ahead of the competition and stay ahead. As long as the leaders keep accessing the most data, it’s virtually impossible for another developer to write a program that’s better, the new software will always be behind because it has to learn everything that the leading programs have already learned and new entrants can never catch up.

We have never seen this before. In the past, we’ve seen newly-developed software overtake older products because the new programs are better. But with machine learning and access to giant lakes of data, companies that become the leaders will retain their leadership and they will own their market segment. That makes right now a unique time in the software business. It’s the equivalent of a software land rush, whoever gets their first, will own the territory and it will difficult to impossible to displace them.

Why Retail

The enormous amount of data that is generated at retail is why it’s the perfect place to prove out artificial intelligence and machine learning. Every time you look at a product in a store, pick it up, put it down or buy it, data is generated that software can analyze. In the past, that information was lost, but now it can be captured and studied. Your actions together with those of billions of other consumers, creates trillions of data points every day for artificial intelligence to learn on and become better. That constant, enormous data creation makes retail an ideal place to prove out artificial intelligence and machine learning and there is likely no better industry for it. A presentation I recently saw said that a particular software company’s programs analyzed 175 years of video every day. Competing programs with less access to data will never catch up to them. Ambarish Gupta, founder and CEO of a new investment fund created to invest in artifical intelligence opportunities, finds that two-thirds of the investments he’s reviewing are retail-related. That’s not a coincidence, retail is the ideal proving ground for software that benefits from analyzing humongous amounts of data. Imagine if retailers had software that told them there was a sold-out item in their store but a box full of the product in the back and staff is needed to unload it. Imagine if software could tell you which product will fit you before you try it on, which breakfast cereal or wine you’d prefer before you buy it, which skincare matches your tone before you apply it. All these technologies exist and they are being tested and implemented right now.

With software development focused on retail, industry performance will improve and the software companies that facilitate the improvement will be highly valuable. If they can continue to access the biggest data lakes, their market positions will be maintained and the value of their equity will skyrocket. And that’s why the biggest opportunity in retail now is in technology, specifically artificial intelligence and machine learning technology. For investors, the biggest opportunities in retail are in retail-related artificial intelligence and machine learning. For retailers, the biggest opportunity to sell more is in giving consumers a better experience based on real learning from trillions of data points. For consumers, the opportunity to more easily get the products they want is an enormous improvement in their experience. We are already seeing evidence of the technology focus on retail. In a survey being released on Thursday of this week, a digital analytics company FullStory, which helps marketers create a better digital experience, found that retail had the highest quality of online experience of any industry tested (airlines had the worst).

Marketers in the past have only been able to guess at what it means when billions of consumers abandon products after looking, or try to buy the same products at the same time. Now, the enormous amount of information generated by that activity is going to generate decisions that improve the experience for all parties. For an industry that sells mostly physical products, it’s strange that the biggest opportunity is invisible, but if retailers are going to succeed in a world where data can be kept on every human action, knowing how consumers act is going to be critical to future success and presents enormous opportunity for retailers and investors who can adapt to this never-before-seen technology.