top of page

How Pattern-based Thinking Gives Companies an Edge

Today’s technology giants, such as Uber and Google, are successful because they introduced something new and innovative to the market, according to conventional wisdom. But Wharton professor of operations, information and decisions Eric K. Clemons thinks that’s too simplistic. Patterns repeat throughout history, and one can find glimpses of today’s new business models in the most successful companies of yore, he says.

Mastering “pattern-based thinking” will help today’s companies get ahead, Clemons argues. He joined the Knowledge@Wharton radio show on SiriusXM to talk about this mindset, which he encapsulates in his book, New Patterns of Power and Profit: A Strategist’s Guide to Competitive Advantage in the Age of Digital Transformation.

An edited transcript of the conversation follows.

Knowledge@Wharton: Why did you decide to write a book about this topic?

Eric K. Clemons: It’s actually a memoir. It’s the history of a great adventure. About mid-1980s, I realized that economics really understood big industry and [Harvard professor] Michael Porter had said just about everything that needed to be said about strategy for traditional manufacturing, transportation and retailing companies. But traditional industry wasn’t where things were happening. Economics had started to look at the power, the value of information.

Economics had started to look at limited or bounded rationality, the fact that we aren’t really perfect economic machines. There was this whole new opportunity to document how information changed strategy in new industries. Given the bully pulpit of being at Wharton, I was able to talk my way into places where I really shouldn’t have been, and I got to do strategy for clients with problems that I would never have thought of myself.

In about 1990, I had a client ask me about the power of online search. Now, [Google founders] Larry [Page] and Sergey [Brin] were probably still wearing short pants at the time, but I did my first paper with Paul Kleindorfer, a regulatory economist [who was at Wharton at the time], on the power of search in 1991. I had a client ask me about whether e-commerce was going to affect the power of traditional retailers, whether it would affect the power of traditional manufacturers, and who would end up owning the channel. It was a great question.

I worked for the president of Lever [Brothers] in the U.S. for years and finally concluded that the power was not going to be with the manufacturers and it was not going to be with traditional brick and mortar retailers, it was going to be with online giants. I designed various scenarios to describe alternative ways the struggle could play out. I didn’t name the scenario we’re in now Jeff Bezos’ Wonderful Adventure. I actually called it Bill Gates’ Wonderful Adventure, but it was clear where the power was going to end up.

I had a client who was, at that point, a four-star [general] and actually the chairman of the Joint Chiefs. He wanted to know if social nets were going to lead to the fall of the government in Tehran. Working with a team of people much smarter than I am in anthropology, we documented the fact that it was much more likely to lead to the fall of [former Egyptian President Hosni] Mubarak than it was to lead to the fall of the Mullahs in Iran.

This is has been such a great adventure. Great men write memoirs. I’m not a great man. But I would say great adventures also deserve a memoir. That’s how I think of this book. I think of this as a great adventurer writing about the emergence of the new economy and writing a strategy guide for all of us.

Knowledge@Wharton: Your book talks about pattern-based thinking. Have you been using it for much of your professional career?

Clemons: It’s a shorthand. It’s an old physicists’ trick. If you’ve seen a problem you’ve never seen before, you try to analyze it in terms of the closest problem you have seen. You focus on essential elements looking for similarities. If you’re predicting the outcome of a car crash, you don’t want to focus on the color of the vehicles — you want to focus on their relative weights and speeds. Similarly, when solving a business problem, you look for the closest match you can find and focus on what’s most important and what’s different between the two problems. You focus on what is essential and find the best available match to the problem you are trying to solve.

“If you’ve seen a problem you’ve never seen before, you try to analyze it in terms of the closest problem you have seen.”

When I was asked to assess Uber for the first time, it looked to me like they had found a way to harness the customer profitability gradient. That is, they had found a way to find the customers they really wanted. They were able to serve customers who were willing to pay for convenience, didn’t want to wait to hail a cruising taxi and were cheaper to serve because you could dispatch a vehicle directly towards them.

Harnessing the customer profitability gradient is also the business model of Capital One. [Co-founder] Nigel Morris said of Capital One, “We’re not a bank, we’re an information-based strategy company, and we surf the customer profitability gradient.” The book really resonated with Nigel’s thinking, which is why he wrote the quote on the back cover of the book. Similarly, the chairman of Uber said, “We’re not a transportation company.” He didn’t say, “We surf the customer profitability gradient.” But the two companies have exactly the same business model.

So, what I’ve been trying to do with my students is teach them how to see what’s essential. When Uber was just introduced, there was no history on the company. You can’t use big data or any historical data, to make predictions. However, you can find something that looks just like Uber. Similarly, when Google was introduced, there was no history on the company. Once again, you can find something just like it, which turned out to be the airline reservation systems of the ’80s.

Knowledge@Wharton: You mentioned that it is harder for some students to truly understand this because they have only lived through the information age. They don’t know what life was like before the internet.

Clemons: I consider myself a digital native, maybe the first digital native. My wife said she was probably the first online computer widow. In 1970, Harvard put a terminal in my bedroom so that I could work 24/7. I’ve been online since long before there really was anything online. What I find interesting is that there is a generation of people who have no idea what life was like before the net, but they also have no idea what’s under the surface of the net. They don’t understand national vulnerabilities. They don’t understand opportunities.

I had a former student say to me, “If you took out half of the Amazon web service data centers on the planet, the world would stop.” That’s probably true. Skynet doesn’t need to blast us. All Skynet needs to do is take down the cloud for a couple of days and we’re all starving. So, you’re right. Teaching pattern recognition is surprisingly difficult when students don’t know what existed before the current online world.

Years ago, I had clients at Microsoft who were looking for a model of Google’s power. I pointed out to them that the best analogy was Sabre and Apollo, the reservation systems at American and United and their lock on search for airfare. They said, “We really should have thought of that.” And I said, “No, it’s a specialized skill. You wouldn’t have thought of it if you weren’t antitrust lawyers.” They said, “We were antitrust lawyers.” And then I said, “Well, you know, there are lots of flavors of antitrust lawyers. You wouldn’t have thought of it if you weren’t an antitrust lawyer at American or United.” And they said, “We were antitrust lawyers at one of the airlines at the time of this litigation.”

I realized that if you change the color of the vehicle you’re looking at, some people can’t immediately see the pattern. Changing from an airline to a search engine may be enough to hide the pattern. So, I find teaching patterns the most challenging and the most powerful part of what I do now.

Knowledge@Wharton: Have we gotten away from pattern-based thinking because data is so readily available now?

Clemons: Absolutely. Business school teaching in 1970 was entirely patterns because we didn’t have any theory and we didn’t have any data, and I was initially not very good at it. The disparaging thing the faculty said about me was, “He’s a really good quant, but he’d never pass his APVs.” That was the Administrative Point of View Exam — the ability to resolve a problem you’ve never seen before using subjective pattern skills rather than algorithms. I think subjective problem-solving and pattern recognition has gone from being the center of the business school curriculum to something that’s almost lost. And I think re-achieving the proper balance is crucial.

“Business school teaching in 1970 was entirely patterns because we didn’t have any theory and we didn’t have any data.”

Knowledge@Wharton: When you talk about patterns, you bring up three important aspects: reframe, recognize and respond. It seems the first two lead to the third.

Clemons: Absolutely. And to be clear, I am a quant. All of my degrees are quant. All of my degrees are from engineering institutions. I am not dissing quantitative thinking, at Wharton or anywhere else. But quantitative thinking can paralyze you.

I remember doing some work with Merrill [Lynch] decades ago in which they were trying to figure out whether to let Mike Bloomberg sell his system to their biggest bank competitors. They were trying to come up with a quantitative number on what their bond trading practice would be worth with Bloomberg selling to the largest banks, and what bond trading at Merrill was worth now. That turns out to be the wrong question, because it was impossible to answer.

The correct question is really very simple, “If we don’t let Mike sell, will somebody duplicate his trading system?” The answer is, “Of course.” Then we can ask, “And if somebody duplicates it, will they get the revenue we would otherwise have gotten?” Again, the answer is, “Of course.” So, now the problem is easy to solve. The world is going to change if we let Mike sell.

The world is going to change in similar ways even if we don’t let Mike sell, because someone else will step in and develop and sell the same system, now that the world has seen that it can be done. The biggest difference between letting Mike sell without restrictions or not is whether we get the revenues from the sale of the trading systems or someone else does.

Here’s the pattern. The world is going to change either way; the status quo is going to vanish. The real question isn’t whether letting Mike sell is better than where we are now; that world is going to be gone. The existence of Mike’s trading systems changed the world forever. The real question is whether letting Mike sell in that new world is better than not letting Mike sell in that new world. Getting the revenue is obviously better than not getting the revenue.

But understanding that the world will have changed, and that we need to ask different questions, is a pattern that we see again and again. It even has a name, it’s called the “Trap of the Vanishing Status Quo.” Under the trap, you’re trying to figure out if life is better if you act in the new world, or if you could somehow remain exactly where you are now. It’s the wrong question. The question you want to ask is, “Is life better if you act, or if you don’t and the world passes you by?” Suddenly, your analysis becomes really simple.

Knowledge@Wharton: If you can understand the pattern historically, you can understand the problem in the current frame, and it becomes easier to solve.

Clemons: It becomes a lot easier. I couldn’t originally find a publisher for the book because the first four or five publishers who looked at it said, “Executives have no patience for history.” I had to reframe the book. I had to make history relevant.

History is not a collection of stories. History is the skill that allows you to recognize a pattern. I need some history to teach the pattern, but that’s really valuable. Once you truly understand the problem, then everything you ever learned about quantitative strategy kicks in. I’m not dissing strategy. But I’m telling you how to do diagnosis first.

Knowledge@Wharton: How does pattern-based thinking affect power and control now, especially when information-based thinking is so prevalent?

Clemons: There is so much information that if you try to analyze all of it you get nowhere. There are occasions when I’ve made analyses without any data at all. For example, in 2006 I predicted for a client that Iran’s Hezbollah proxies in Lebanon would launch missile attacks on Israel.

I had no data at all to support the suggestion, but it just made sense. It created such respect for Iranian intransigence on the Arab street that it was almost impossible for any Arab government to oppose a Shia bomb. When the attack occurred several months later, my clients said, “How did you see that coming?” The answer was, “Given the expected result, that’s what I would have done,” not, “That’s what the data said.”

If you can understand the context in which you operate, you understand what your competitors should want to do, and then you understand what you should want to do. The guy who understands first is the one who moves correctly first, and the guy who moves correctly first often wins.

“After the Industrial Revolution, we needed protections against externalities, actions taken by a company that could affect others.”

Knowledge@Wharton: In the book, you explain the third-party payer system, which I think plays into what you’re talking about.

Clemons: Third-party payer systems scare me. They always start off safe for everybody. Google was the fastest way to find anything. We’ve all become Google addicts, myself included. Corporations absolutely have to be found. Now, if all of us searchers paid to use Google, the price we paid would be limited by the reduction in search time, how much effort it saves us.

But if none of us can find what we’re looking for without Google, then the amount Google can charge is not limited by the value to the searcher — it’s limited by the full value of the sale to the company that needs to be found. Paying to be found is not nearly as valuable as paying to not being not found.

“The world is going to change either way; the status quo is going to vanish.”

I was working with a hotel chain that for obvious reasons won’t let me use their name for fear of angering Google. We did experiments in which we occasionally refused to purchase our own name as a search term. We found that every dollar we saved by not paying for our name cost us between $30 and $40 in lost revenue, so we paid whatever we were told the cost of our name would be. That’s pretty scary.

Even if customers were explicitly looking for us, Google could send them to a different website and we lost revenue. In the worst case, we lost the full value of the sale. If the searcher, the first party, pays, the most Google can charge is the value of convenience. But if the hotel or merchant, the third party, pays, it doesn’t matter how much Google charges as long as the searcher is kept happy.

Sellers might be unhappy, they might even be very unhappy, but that doesn’t matter either. As long as buyers stay with Google, sellers have no choice. They have to continue to pay to not be not found. The third party is paying, and since they have no alternative, that’s why I call these systems mandatory participation systems. Again, the price charged is limited only by the value of the sale to the seller. It has nothing to do with the value the system creates for the searcher.

We learned from Sabre and Apollo that if you were listed in Sabre but not Apollo, you went bankrupt. If you were listed in Apollo but not Sabre, you went bankrupt. If you’re easy to find in Bing but not Google, you die.

Google’s ability to charge is not limited by what the first payer, me, first party thinks it’s worth, which is the value of convenience. It’s limited solely by the value of the sale, the value of not being not found, which is enormous.

When we wrote about this in 1991, it didn’t seem like a big deal other than theoretically interesting. But it’s now the airline reservations systems model on steroids. It’s so big and it’s so powerful and the revenue is so all-inclusive that it may be the most important emerging business model on the planet.

Knowledge@Wharton: Towards the end of the book, you discuss the impact of these types of decisions on our society as well. Can you explain?

Clemons: As I’ve gotten older, I have become such an avid supporter of Western democracy and all of the property rights and business rights that that entails. Not because it’s perfect, but because it’s better than anything else I’ve seen. So, I am frightened when I see business models that are potentially quite disruptive to our shared values as a society. I don’t just mean wealth inequality, though at some point that becomes intolerable, I mean the marketization of everything.

I’ve had Uber drivers who’ve told me with joy that they were buying buildings in neighborhoods that had been predominantly working class and ethnic minority, and gentrifying them and turning them into Airbnb properties. They were forcing out the people who have previously lived there and changing the neighborhood.

It’s not Airbnb’s responsibility to set social policy, but somebody has to set social policy in the sharing economy. After the Industrial Revolution, we needed protections against externalities, actions taken by a company that could affect others. We needed zoning laws and regulations to control pollution. What do we need to control externalities caused by giant platform companies today?

If I list the things I’m worried about, and this is going to be a little over the top, one of them would be that technology changes governments. I don’t mean it changes regimes; it changes the entire nature of government. After the Peace of Westphalia in the middle 1600s, gunpowder, artillery and the deployment of permanent standing armies created the modern nation state and pretty much determined the map of Europe more or less as we see it today. How will modern technology change the nature of nation state now?

What about the unprecedented power of Facebook to destabilize the government, as exemplified by fake news, manipulation of the 2016 presidential election or the Brexit referendum? What about the power of Google to determine what we find when we try to learn about candidates or issues? Surely Google, Twitter and Facebook are going to maximize returns to their shareholders as long as their actions are not explicitly illegal; it may now be appropriate for governments to limit the behavior of these firms in order to protect citizens, the future of democracy and indeed the future of the nation state.

bottom of page