(Welcome to my series on an “Intelligence Preparation of the “Streaming Wars” Battlefield”. Combining my experience as a former Army intelligence officer and streaming video strategy planner, I’m applying a military planning framework to the “streaming wars” to explain where entertainment is right now, and where I think it is going. Read the rest of the series through these links:
Part I – Define the Battlefield
Defining the Area of Operations, Interest and Influence in the Streaming Wars
Unrolling the Map – The Video Value Web…Explained
Aggreggedon: The Key Terrain of the Streaming Wars is Bundling
This is probably the most popular image for business school students about Amazon. Heck, anyone describing Amazon has probably used this image.
If we’re supposed to be neutral observers of businesses, you can’t help but notice after a moment of reflection how insanely positive this take is. Man, Jeff Bezos can really sell his positive vision and have it repeated universally.
If you were really cynical—hey, I am—what would the pessimistic version of this flywheel look like? The “Flywheel of Evil” if you will…
What changed? Well, first, the idea that you “sell more things” is great, but if you lose money on every transaction, that’s “sub-optimal” in business speak. Or bad in human speak. And Amazon does in many cases.
To fund these losses, you need to start a really successful company that is totally unrelated to your retail business or its membership program, which is where Amazon Web Services comes in. There’s an alternate history where an Amazon without AWS (cloud computing) doesn’t take over retail because it doesn’t have a cash flow engine driving its growth. (In that timeline, Ebay becomes our overlords.)
Even more potent, though, is combining already low prices with Amazon’s decades long refusal to pay local taxes. Could you point to the continued imprisonment of poor Americans to online companies not paying local taxes? Maybe! (As local tax bases erode, some communities turned to police forces to extract rents, like in Ferguson, Missouri. Seem relevant to our current times?) Amazon does pay some local taxes—now—but only after it became an advantage to them in furthering their monopoly power.
Now that it has this “flywheel” rolling, Amazon uses its size to both crush new entrants who want to compete and to punish suppliers, capturing all the value from their product creations.
Which flywheel is “right”, then? Well, both actually. Both describe valuable methods for how Amazon grew to the size it did. Some of those methods were good for customers; some were bad for society. You can’t tell their story without both.
What’s the lesson? Flywheels are simple whereas reality is complicated. As tools, flywheels are fairly inexact. They’re not even really tools, but narrative devices we use to help make sense of a complicated world. In other words, a “heuristic”. As behavioral economists like Kahneman and Tversky taught us, heuristics are useful, but can carry pitfalls if we aren’t careful.
What’s the point for the streaming wars? Well video has become a spoke on multiple company’s supposed “flywheels”. Everyone from Disney to Amazon, but most critically Apple last fall. Whether or not these were actual flywheels was less important than merely invoking the term and using it to justify nearly any amount of spending.
Let’s call this another key piece of “terrain” in the streaming wars. The “Forest of Flywheels” if you will. The problem is the business and entertainment press has been fairly sloppy with our language when it comes these types of endeavors. Due to this sloppiness, we’ve allowed a lot of companies to launch video because they’ll “lose money on video to make money on X”.
Today, I’ll explain the key terms. In my next article I’ll critique deficit-financing in particular. And then I’ll finish it off with an analysis of some of these business models to show their potential strengths and weaknesses.
– Flywheels are the most overused term in business, and it’s important to know what different terms mean.
– Ecosystem is probably the most commonly confused term with flywheel. Ecosystems are also rare.
– A true flywheel is a self-perpetuating cycle of growth that is incredibly rare in practice.
– As such, in pursuit of flywheels, we’ve seen many digital players launch money-losing video efforts. I call these “deficit-financed business units”. And they’re one of the biggest factors in the streaming wars.
Defining Traditional Business Strategy Terms
You’ve read articles bemoaning jargon in the workplace. (This New York Magazine piece is the latest in hundreds on the subject.) Even I just denigrated “sub-optimal” above, a term I really don’t like. Still, I don’t take that extreme of a position on business nomenclature. Often, jargon really does have a role in explaining new concepts.
The problem comes in overuse. That’s what is currently happening with “flywheel”. It’s almost become synonymous with “successful business”. But it’s much more specific than that.
So let’s define our terms, so we can better understand what is and is not a flywheel.
It turns out if you want to stymie business school students, just ask them “what is a business model?” Indeed, they’re taking classes called “Strategy and Business Models”, but answering, “What is a business model?” can stump them. I’ve seen it.
At its most basic, a business model is a plan or process to make a good or service and sell it for more than it costs to make. Make a widget for $1, market it for $1 and sell it for $3. Or replace widget with service. The model is how you make money. On a financial statement, this is usually called the income statement. When I build a “model” for this website, that’s usually what I’m building.
How do business models relate to flywheels? Well, you can have a successful business model that isn’t a flywheel! It’s just a good business. In the olden days, you would have probably described the dividend producing stocks as just good businesses. They don’t have huge growth prospects, but they still generate a return on investment. Cable companies in the 2000s fit this bill. They had good business models, but were absolutely not flywheels.
Where it gets complicated is usually a given company is actually a collection of many business models. Arguably for every product they sell. Or you have distinct models for different business units in the same conglomerate. Which is actually a good transition to our next definition.
Most companies on the S&P 500 aren’t just one business, but multiple types of businesses lumped together. This is the reality for most conglomerated businesses. When analyzing a compnay, it’s key to differentiate between its overall success and the success of its various pieces.
Amazon is a perfect example here. Retail is one business unit. But then it also has media businesses from live streaming to streaming to music. Then it also sells devices like Amazon Echo. Oh, and it has Whole Foods groceries too.
And then there is the cloud computing (AWS). Which I called out above. And it’s worth noting just how distinct that wildly financially successful enterprise is from the rest of Amazon’s consumer-focused retail efforts. It’s a business-to-business service that is powered by lots of fixed capital expenditure data warehouses. It barely relates. Yet, it’s part of Amazon.
How do business units relate to flywheels? Well, flywheels often fail to take into account entire business units. Take the Amazon flywheel of success…it totally ignores AWS! For years Amazon survived because it had an incredibly high margin business in cloud computing that could provide necessary capital that enabled Amazon to continue building its retail business. This also kept Wall Street happy.
That makes the Bezos flywheel not just wrong, but almost negligently wrong.
It’s business malpractice to point out that a flywheel helped Amazon to succeed if you don’t include AWS’s role in propping up the balance sheet!
I would add, many of the “flywheel” charts you see out there are often just describing a company with multiple business units. (I’ve seen this with Disney and Epic Games.) Every business can benefit from owning multiple business units, from lowering costs or providing learnings. That used to be called “synergy”. Now we call them “flywheels”.
Like flywheels, “ecosystem” is often overused and confused.
So what is it? I prefer a simple use of the term. Unlike a natural ecosystem—which is a complex set of interactions between organisms that needs to be in harmony to work—a technological ecosystem means a hardware or operating system platform with connected applications. Usually, the owner of the ecosystem puts up walls to keep the ecosystem closed to unwanted applications.
Apple is a good example of a company with multiple business units all dedicated to reinforcing the same “ecosystem”. Laptops and computers run one type of operating system, that syncs seamlessly with their mobile operating system. Together, these operating systems run a number of applications. Now, they are synced across the cloud. Most Apple business units reinforce the Apple ecosystem.
I prefer this limited usage to the broadening usage which gets to the point where it describes any company with multiple business units that sometimes work together as an ecosystem. To pick on them both again, I’ve seen Disney referred to its multiple business units as an “ecosystem”. But it’s not: it’s a company with multiple business units that sometimes work together. Same with Epic Games. That isn’t an ecosystem, but several business units that can work well together.
Amazon has worked to build consumer ecosystems, to mixed results (Fire Phone). The Alexa is a start, but most folks still have a smart phone. You could argue Amazon’s retail shop is a seller’s ecoystem, but there is a different term that explains it even better…
Demand-Side Increasing Returns (Or Network Effects)
The words outside the parenthesis are the business school terminology. But we all just call it “network effects”. And yes, this all gets lumped in with Ben Thompson’s “Aggregation Theory”, which is fairly similar.
What it means is that some services or businesses get more useful as more people use them. The obvious example is Facebook. When it was just a service on Harvard’s campus, it had limited use to only Harvard students. Then it expanded to all college students. But this didn’t increase its value linearly, but exponentially. A service with every student is multiples more valuable than one with just one or two colleges.
Overcoming network effects are a huge hurdle new entrants. The challenge to dethroning Facebook even to this day is to get as many customers as Facebook. If you only one tenth the size, you aren’t one tenth as valuable, but one-hundredth. That’s the power of network effects.
How does this apply to flywheels? Well, a business can be successful because it has network effects, but that doesn’t mean it has a “flywheel” of success. Merely that it successfully established a business with network effects and can simply keep that going with minimal investments. I’d put Facebook into the category. Sure, you can make a customer flywheel about it, but really it’s just a popular network that grew quickly. SnapChat and Instagram saw the same growth, and they don’t all have “flywheels”. Even Amazon is more driven by network effects than the flywheel, especially in the Amazon Marketplace.
Be ware! Not every good business is driven by network effects. Sometimes great, fast growing businesses have another force at work…
You haven’t heard this one, had you? I threw it in because it explains a lot of business models better than flywheels. The definition is a system where performance is limited by the weakest unit in the system. (Which happens for a lot of supposed flywheels.) The upside comes when the chain-link system is unique. As said by Richard Rumelt in Good Strategy/Bad Strategy, if a chain-link system is designed to work together, competitor’s often struggle to match it, since they would need to implement the entire system, not just one part
A good example of this is Walmart. Walmart doesn’t have network effects because it’s not like the additional customers make the service more valuable. Instead, what it has is a series of decisions made by its founder that work together to reinforce each other. It started with many distribution centers and centralized planning of products. It’s competitors allowed individual store autonomy, and as a result Walmart could reach smaller markets than K-Mart or Target.
Ikea has a similarly chain-linked, well-designed system: it owns all the goods it ships, it produces in bulk and as a result has low prices. Most other furniture stores and manufacturers can’t emulate it without copying the business model from the ground up, which they aren’t designed to do. (They don’t own manufacturing.)
How does this relate to flywheels? Well again some businesses that are just cleverly designed chain linked systems are described as flywheels. And since we have gotten this far without mentioning them, how about saying this could apply to Netflix?
I’ve seen Netflix’s success credited to a flywheel, ecosystem or network effects. Instead, it’s a well-designed chain-linked system.
The system is designed to flow together: have an on-demand system, that provides data on user behavior that enables better platform optimization. Indeed, the biggest supposed driver of a Netflix flywheel is using data to make better TV shows. But Netflix’s own behavior has shown that data doesn’t help. Instead, they make TV shows in bulk—more than ever produced at one studio. It’s the opposite of more data driving better decision making which is what is their most common flywheel. Further, the history of Netflix isn’t one of a flywheel, because the biggest source of growth was from cheap library content, which has disappeared, meaning the flywheel would have lost one of its spokes. (The cynic would add that Wall Street also allowed continued free cash flow losses.)
The “well-designed chain-link system” explanation for Netflix also explains why legacy media has had such a hard time emulating their business. See, Netflix was created from the ground up to operate the way it does. Traditional entertainment was set up in an opposite way. This is just like Circle K or Target trying to emulate Walmart or Amazon. It didn’t just mean changing how they did one piece of they business, but changing their entire company. And flywheels have nothing to do with it.
Finally, What is a Flywheel?
The term was coined by Jim Collins. A business model or ecosystem or chain-linked system isn’t a flywheel just because it exists. A flywheel is a set of reinforcing business decisions that become self-perpetuating. It’s hard to find definitions of it because it is more of a metaphor than actual concept. Here’s the original metaphor Collin’s own website:
The best definition I found was this one from Medium:
This should provide several key insights related to all our discussion above. First, the flywheel has to reinforce at each step. Not just tangentially, but truly provide value back to the next step. Many businesses, I would argue, aren’t wheels at all, but just a line of value going downward, potentially ending with “customers are happier about your brand”.
Next, the flywheel stores up energy, meaning at some point it becomes self-perpetuating. This is the hope of many technology unicorns is that they would keep losing money until some point that the wheel clicks on like Amazon or Facebook. (How many companies point out that Amazon lost lots of money initially?)
Further, a monopoly is not a flywheel. Instead, it’s a company that has achieved dominant market position and uses that to extract rents. That’s not reinforcing because of business dynamics, but simply because size crushes competitors. Many so-called flywheels are actually just good business models, sometimes with network effects, but they’ve continued because of monopoly dynamics. Don’t confuse the two!
Finally, given all these hurdles it should be clear that true flywheels are rare. Which is what Collins saw as the appeal of them. They are valuable because they are rare and rare because they are so valuable. Our persistence on calling everything a flywheel has cheapened the term.
The Key Insight for the Streaming Wars
This is a long, winding road of definitions to end up back at our topic of this series, the streaming wars. One of the biggest trends in the streaming wars is that big tech companies are launching streaming video services because they see them as reinforcing a flywheel or ecosystem. Many authors wrote about Apple TV+’s launch as if it were a revelation that its goal wasn’t to make money, but to sell more iPhones or drive more subscription revenue.
The risk is that if the video serves another purpose besides making money in and on its own that a company has essentially set up a money-losing business. I call this deficit financing.
But worse, if the new video services doesn’t actually power a flywheel or bring more customers to the ecosystem, it’s just another business unit. (See that’s why I defined the term above.) So what do you call this?
Deficit-financed business units.
And I’ll make this currently controversial, but formerly common sense claim: losing money is bad. Making money is good. Even in video and even in flywheels. It’s not that all deficit-financing is bad, but it is much riskier than the alternative, which is NPV-positive business units. But to explain why, I’ll need another few thousand words.