Your profit is governed less by your product than by the price of its complement — the thing customers must buy alongside it for yours to be worth having. When that complement gets cheaper, demand for your product rises; when it gets dearer, your demand falls. So one of the highest-leverage moves in strategy is deliberately counterintuitive: drive the price of your complement toward zero, on purpose, often by giving it away, because a cheap complement is a demand pump aimed straight at the layer you actually own. Most sophisticated operators know the phrase and almost none run the play deliberately. That gap is the opportunity.
The principle is Joel Spolsky's — "smart companies try to commoditize their products' complements" — and it sits on a piece of first-year economics that is more load-bearing than it looks. Two goods are complements when a fall in the price of one raises demand for the other. Cars and gasoline. Consoles and games. Phones and apps. Grill and charcoal. The demand curve for your product is not a fixed thing you inherit; it shifts right every time a complement gets cheaper. You don't have to build a better product to sell more of it. You can make the thing next to it cheap.
The mechanism, walked slowly
Take the console. Sony and Microsoft sell hardware near or below cost — sometimes at an outright loss per unit — and the reason is not generosity or a loss-leader gimmick. The console is a complement to games, and games are where the margin lives, through licensing and a platform cut on every title sold. Cheap hardware raises the installed base; a larger installed base raises game demand and developer participation; more games raise hardware demand again. The company deliberately commoditizes the box so that value accretes to the software layer it taxes. The box is a toll booth priced to maximize traffic, not to earn on the booth.
Now flip the same logic and watch who is commoditizing whom. In the PC era, Microsoft wanted hardware cheap and interchangeable. Commodity, competing PC makers meant abundant, low-priced boxes running Windows — and every cheap box sold more of the operating system and Office, the layers Microsoft controlled and no one else could supply. Intel wanted the reverse-image version: it wanted the OS and the box commoditized so that all the differentiation, and the margin, lived in the chip. Two giants, each quietly working to commoditize the other's layer, both correct, and the PC assembler in the middle — Dell, Compaq, Gateway — got ground into a few points of margin between them. That squeezed assembler is the position you never want to occupy, and half of what follows is learning to recognize when it is you.
The cleanest way to make a complement cheap is to make it free and abundant, which usually means open-sourcing it or subsidizing it directly. Google funds Chrome and Android and gives them away because a cheaper, faster, more universal mobile-and-web layer is a complement to search and advertising — the layers Google monetizes. Every marginal hour of cheap internet access is demand poured into the ad auction. IBM poured hundreds of millions into Linux to commoditize the operating system beneath its consulting and hardware business: if the OS is free, the customer's budget is freed up to spend on the layer IBM sells. None of these are charity. They are demand pumps, and the pump is aimed with precision at a layer the funder controls and can defend.
The four questions the principle forces
The insight is only useful if it becomes a diagnostic. Four questions, in order.
First: what is actually complementary to my product? Not what I bundle, not what I upsell — what must the customer also have for my thing to deliver value, such that if its price fell, they'd buy more of mine. The formal handle is cross-price elasticity, but you don't need the equation, you need the honesty to name the real dependency. For a streaming service, cheap bandwidth and cheap smart-TV hardware are complements. For an electric car, cheap charging and cheap electricity are complements — which is exactly why charging networks are strategic terrain, not a utility afterthought. At Kommerce, running cash-on-delivery commerce in low-trust markets, our demand is governed less by connectivity than by the cost and density of last-mile logistics and cash reconciliation: when a courier network gets denser and cheaper to settle with, demand for the commerce layer above it rises even when our product hasn't changed a line. Map them all, not just the obvious one.
Second: who profits if that complement is cheap versus dear? Every complement has someone rooting for its price to fall and someone rooting for it to rise, and their interests are usually invisible until you draw the map. When you draw it, the strategy of the whole industry snaps into focus — you can see which giant is subsidizing which layer and why, and you can predict their next move from their margin structure rather than their press releases.
Third: can I commoditize it, and if I do, where does the shifted demand land? This is where most people misplay the principle. Commoditizing a complement destroys profit in that layer — that's the point — but the destroyed profit only reappears in your pocket if you own a layer scarce and defensible enough to catch it. Open-source the complement and the demand shifts, yes, but if your own layer is equally contestable, you have detonated industry profit and captured none of the blast. The move is only rational when you can name, specifically, the non-commoditized layer the demand flows into and why you hold it. That controlled layer is the actual moat. Commoditizing the complement is how you widen the moat; it is not itself the thing keeping you alive. This is why "we have network effects" is so often a misread of where the defense really sits — advantage is a metabolism you run, not a wall you inherit, and the layer worth running your metabolism on is the one you have deliberately kept scarce while everything adjacent goes cheap.
Fourth, and the one people flinch from: am I someone else's complement being commoditized right now? This is the scariest position in business and the one your own optimism hides from you. If a well-capitalized player above or below you is funding a free or open version of what you sell, your margin is not eroding by accident. It is being competed to zero on purpose, as a deliberate act, to enrich the layer that captures the demand you're about to release. The assembler between Microsoft and Intel didn't lose to a better assembler. It lost because two neighbors needed it cheap.
The signature of being zeroed out
Three signs, and when they appear together you should treat it as a fire, not a weather pattern. Your product starts getting described — by customers, by analysts, by your own salespeople — as "just a feature" of something larger. A serious player is funding a free or open substitute for exactly what you charge for. And your pricing power is slipping even as your usage climbs, which is the tell that separates being commoditized from ordinary competition: in normal competition, growth restores some pricing power; when you're a complement being zeroed, growth just makes you a more attractive thing to make free.
The escape is not to fight the commoditizer on price — you will lose, because their entire economic purpose is to lose money in your layer. You move. Up into a layer that is scarcer, or down into one that is harder to replicate, or sideways into genuine neutrality so that no single commoditizer can zero you without handing the gain to a rival. The one move that never works is standing still and defending the margin in the layer someone richer than you has decided must be free.
The AI-era version, marked as interpretation
Here the map is being redrawn live, so read this as a live reading rather than a settled fact. For most of the last few years the strategic story was that foundation models were commoditizing the software that sat above them — the thin wrapper apps, the "we built a UI on top of an API" companies, whose entire product could be swallowed by the next model release. Being a complement to a model that keeps getting cheaper and more capable is a brutal seat: your value proposition is exactly what the layer below you is racing to absorb, and the price of the complement (tokens) is falling on a curve that would make a semiconductor engineer blush. Falling token prices are, for an application, a cheap complement — great for demand, until you notice the model provider can walk up into your layer at will.
The arrow may be pointing the other way too, though, and this is the part I hold loosely. Model providers appear to want the application and tooling layer above them to be abundant and cheap — they subsidize developer ecosystems, open frameworks, cheap tokens — because a thriving, low-cost layer of apps is a complement to model consumption, which is what they monetize. Cheap apps pump demand for tokens. Simultaneously, several labs are releasing open-weight models, which reads as an attempt to commoditize rival models — to make the model layer itself cheap and abundant so value has to accrue somewhere the releaser holds, whether that's cloud, tooling, or distribution. Open-sourcing the model is the IBM-and-Linux move pointed at your competitor's core product. Whoever is currently deciding that some layer of this stack "should be free" is telling you, in the plainest possible language, which layer they intend to get rich on. Watch the subsidies, not the mission statements.
What to actually do this week
Draw two columns. In the first, list every complement to your product — everything a customer must also have for yours to create value, however far from your P&L it sits. In the second, for each one, write who benefits when its price falls and who benefits when it rises, and mark which of those parties has the capital to act on that interest. Two things fall out of that page. You will find at least one complement you could make cheaper — through open-sourcing, standardizing, subsidizing, or simply refusing to let it stay a chokepoint — that would shift demand into a layer you control, and that is your offensive move. And you will find whether you yourself sit in someone's first column, being quietly readied for zero, which is your defensive one.
The waste you sink into commoditizing a complement, like the best marketing, looks irrational right up until you see the layer it's feeding — and the surplus abundance you create in one layer is frequently the raw material for a second business entirely, which is the argument for learning to monetize the byproduct rather than the product. Both moves start from the same refusal: the refusal to believe your profit lives inside the thing you happen to sell.
Cheap is a weapon. The only question is whose hand it's in.