Where to Play: Lessons from Amazon’s Full-Stack Strategy

This week, Amazon announced that it’s now selling logistics as a service. The company currently provides warehousing, freight, fulfillment, and delivery functions to sellers on its online marketplace. Amazon Supply Chain Services will now offer these same services to other businesses as well, putting it into direct competition with FedEx and UPS. Shares of those carriers fell nearly 10% on the news. It’s never a good thing when one of your largest customers goes into business against you.

Commentators noted that the expansion into logistics is similar to the company’s launch of Amazon Web Services. In the early 2000s, Amazon was building out massive computing infrastructure to support its own retail business. Along the way, the company realized that much of that infrastructure wasn’t unique to retail. Every digital business needs the same basic building blocks. So, Amazon began offering those capabilities to others, first in small ways, and then more formally as AWS. What started as an internal necessity became a product in its own right, and eventually one of Amazon’s most profitable businesses.

That kind of move rarely happens outside the technology sector. By and large, healthcare companies look for their next move within healthcare. Entertainment companies look for new ideas in entertainment. Banks tend to grow by acquiring more banking rather than rethinking the system around them. And yet, tech behemoths like Amazon, Google, and Apple seem to have no fears about playing in those other spaces. That’s partly due to how much cash they have. But it’s also a result of a different view of strategy. Large tech companies are always exploring ways to provide the other layers needed to deliver value to end customers. They’re constantly looking at the full stack.

Full Stack Strategy

The idea of a “full stack” has deep roots in how technologists think. Early computer systems were a tangle of interdependent parts that were hard to manage. Those parts included everything from hardware to memory to program instructions. A change in one place could ripple unpredictably through the rest of the system.

To make sense of that complexity, engineers started breaking the system into layers, each with a distinct role. Each layer could pass information to the layers above and below it without needing to understand how the other layers worked internally. So long as the overall system architecture stayed the same, individual layers could be swapped out or improved.

Over time, this idea of a system architecture got applied to software. By the 1990s, developers thought about software as having a database layer that stored information, application layers that did the computation, and an interface layer that presented information to the user.

Some engineers became specialists in a particular layer. Others saw themselves as “full-stack engineers,” the kind of people who could work across many layers.

That way of thinking has shaped how technology companies approach strategy. When they think about where to play, they identify their place in the stack. And they think about how their layer connects with other layers or obviates them altogether.

Companies outside the tech sector tend to approach strategy differently. When they consider where to play, they think about new markets, new customer segments, or new categories. The closest a non-tech company gets to examining its stack is when it considers where it plays in the value chain. And it rarely seeks to become a “full-chain player.” That’s because value chains in the real world often span wildly different domains and competencies. The know-how required for open-pit graphite mining is a world away from selling pencils.

Amazon and other tech companies operate from a different premise. To them, the boundaries of an industry aren’t fixed. They’re design choices. If the system doesn’t work as well as it should, the question isn’t how to optimize your position within it. The question is whether you should reshape the system itself.

Amazon started as a retailer but quickly built a marketplace to expand its selection. It invested in fulfillment because relying on others limited its speed and reliability. It moved into last-mile delivery for the same reason. Each step was driven by a problem inside the system, and, interestingly, most steps also provided value to other players in the stack.

What looks like expansion from outside the company often feels like a necessity from the inside. The company keeps encountering constraints, and instead of working around them, it replaces them. Over time, that leads to it taking over more layers in the stack. Apple launched its own stores because it couldn’t get adequate attention at CompUSA. Carrying that thinking to its logical extreme, the company now designs processors, creates phones, builds software, sells apps, distributes products, and runs retail stores.

Any first-year business student will ask why this isn’t just old-style vertical integration. Vertical integration and full-stack expansion both end with a company owning more of its own system. The difference is in what gets built along the way. Vertical integration is about bringing external suppliers in-house to reduce costs and increase control. Full-stack thinking does that, too, but it also looks for layers where the same infrastructure can serve other players. Sometimes that means taking something you built for yourself and selling it to others, the way Amazon now sells logistics to companies that have nothing to do with its marketplace. Sometimes it goes further, and the infrastructure itself becomes a platform that other companies build their own businesses on, as Amazon did with AWS. It’s all about creating new revenue from new customers.

Many of the layers that Apple develops serve as platforms for other players in the stack. Apple Stores sell a curated selection of non-Apple products, too. The vast majority of apps in its App Store are made by other companies. As a result, much of the value that people get from their iPhones comes from Uber, Spotify, and a host of other players that exist in the Application Layer of Apple’s stack. That’s more than vertical integration. That’s what happens when you bring an engineer’s mindset to thinking about strategy.

Full Stacks in the Non-Tech World

While rarer, this kind of approach to strategy does show up outside tech companies. And no better example exists than Reliance Industries.

When Dhirubhai Ambani returned to India from the Middle East in the late 1950s, he didn’t set out to build one of the world’s largest energy companies. He began as a trader, importing polyester yarn and selling it into India’s growing textile sector. Before long, Ambani moved into manufacturing, launching a textile business of his own, and then a consumer-facing brand that would eventually own its own retail shops. At that stage, Reliance looked like a straightforward play in finished goods, competing on quality, distribution, and brand.

But Ambani saw something others didn’t. The economics of the clothing industry were dependent on the cost and availability of textile fibers. The cost of synthetic fibers was, in turn, dependent on the cost of petrochemicals. In tech terms, these were layers further down in the stack.

Reliance began moving further down through the system. First, it invested in making polyester filament yarn. Then it built the chemical processing plants that cracked naphtha to make polyester. Over decades, Reliance steadily built capabilities that most textile companies would never have considered. It then turned around and sold its goods to other textile producers.

By the 1990s, Reliance had expanded into large-scale petrochemical plants that sold gasoline on the open market while producing the raw materials that fed its own manufacturing. What had started as a textiles business was becoming an integrated stack of multiple businesses. In 1999, the company commissioned its first oil refinery, enabling Reliance to control a system that started with crude oil and ended with men’s suits. What started as backward integration to feed Reliance’s own polyester business is now one of the world’s largest export operations for refined petroleum products, selling gasoline, diesel, and jet fuel into more than 100 countries.

Three Factors for Evaluating Expansion

To be sure, both Amazon and Reliance have been able to deploy a full-stack strategy because they grew in environments that favored that kind of expansion. In the latter half of the 20th century, India was a protected playground for a small number of oligarch families. Owning a business made it easier to start another one. Conversely, American e-commerce in the last thirty years offered tech players the power of network effects and a laissez-faire regulatory environment.

Also, there’s a bit of survivorship bias here. We’re only considering the folks who did it well. While the results can be transformative, they can also be catastrophic. Too often, they end up as just a painful distraction. The difference between a future-focused move and a costly misstep often comes down to a few underlying factors.

Thinking about your business as part of a stack starts with a definition of who you’re trying to create value for. Business expansions that serve the same customer as before have a higher likelihood of success. In some cases, though, your biggest customers turn out to be those other players in the stack. Amazon’s retail customers didn’t need a logistics service. But Amazon did, and so did the millions of sellers who rely on its platform. The need wasn’t with end consumers, but it existed in the ecosystem.

The next factor to consider is your own capabilities. Those capabilities exist in two forms: your assets and your abilities. Assets are what you own. Abilities are what you’re good at. Many companies start with their assets and try to stretch from there, asking how they can extend what they already have into adjacent markets. But as a rule, abilities matter more than assets. And developing a full stack often demands building assets you don’t yet have.

Amazon didn’t begin with a logistics network. It built one over time, and it built it at enormous cost. But that investment was grounded in world-class abilities. The company already knew how to design and operate complex systems that combine software, physical infrastructure, and large-scale operations. The asset came later. The ability came first.

The third factor is whether the move reinforces the core. Many expansions end up weakening the business they were meant to support. In Amazon’s case, the opposite dynamic is at work. As more volume flows through its network, costs come down, delivery speeds improve, and the Prime experience becomes more compelling. The new business doesn’t sit alongside the core. It strengthens it.

When these three elements come together—a real need, an underlying ability, and a reinforcing effect on the core—the result tends to be more than just a new revenue stream. It becomes a coherent extension of the business itself.

The Stakes Are Rising

Industries like financial services and healthcare are shifting dramatically as tech companies expand their place in the stack. Businesses that were built to dominate a single part of the value chain may find themselves operating within someone else’s stack.

Companies like UPS and FedEx aren’t passive observers in this story. They are making deliberate choices about where to focus, how to invest, and which relationships to prioritize. When one company begins to define the system itself, everyone else has to decide how they relate to that system. In January 2025, UPS announced that it was intentionally reducing its revenue from Amazon by 50%. That makes UPS less dependent on one of its largest customers, but it doesn’t change where it plays. That will likely come from market segments like healthcare—segments that are currently outside Amazon’s stack.

Both FedEx and UPS face important strategic choices in the months ahead. Do they try to move up the stack, closer to the customer and the data that drives demand? Do they double down on their role within a specific layer, becoming even more indispensable as part of a larger ecosystem? Or do they find new ways to participate in a system they no longer control?

Full-stack thinking isn’t reserved for tech giants. It starts with a different way of seeing your own business. Revisit what you’re truly good at, where value is created around you, and which parts of the system you’ve been taking for granted. From there, the path opens up. You don’t have to own every layer. But you can begin to understand the stack you’re in and decide, deliberately, where you need to play next.