What Clayton Christensen Would Tell the SaaS Industry Right Now

The Innovator's Dilemma Has Entered the Chat

Something is happening in the SaaS industry that we've seen before. Not in software—in steel, in telecommunications, in photography, in data storage. Every time it happens, the incumbents see it coming. They have better technology, deeper pockets, stronger customer relationships, and smarter people. And almost every time, they lose anyway.

The cost of building and delivering software is dropping by 80–90%. AI isn't just making existing development faster—it's enabling entirely new architectures with fundamentally different cost structures. If you're running a SaaS company right now, there is exactly one person whose work you should be studying, and he's been dead since 2020.

His name was Clayton Christensen, and he wrote the playbook for what's about to happen to you.

The Professor Who Explained Why Smart Companies Fail

Clayton Christensen was a Harvard Business School professor who, in 1997, published "The Innovator's Dilemma." The book asked a deceptively simple question: why do well-managed, customer-focused, technologically sophisticated companies consistently lose to scrappy upstarts with inferior products?

His answer was uncomfortable. These companies didn't fail because of bad management. They failed because of good management. They listened to their best customers. They invested in higher margins. They made rational decisions at every step. And those rational decisions, compounded over time, led them straight off a cliff.

Christensen didn't just theorize about this. He documented it across dozens of industries with meticulous data. He showed that disruption follows a pattern so consistent it's almost mechanical. And that pattern is now playing out in enterprise software with uncomfortable precision.

The Framework: Why Rational Decisions Kill Companies

Christensen's framework centers on a distinction between "sustaining" and "disruptive" innovation. Sustaining innovations make existing products better along the dimensions that existing customers already value. Disruptive innovations introduce products that are initially worse on those dimensions—but dramatically cheaper, simpler, or more accessible.

Here's the critical mechanism: when a disruptive technology appears, it almost always looks like a toy. It can't serve the incumbent's most profitable customers. A rational manager, looking at the numbers, correctly concludes that there's no business case for pursuing it. The margins are too thin. The customers are too small. The technology isn't mature enough.

So the incumbent does the smart thing and ignores it. Or hedges with a small internal team that gets starved of resources because the core business is still growing.

Meanwhile, the disruptor improves. Not by trying to match the incumbent feature-for-feature, but by iterating rapidly in markets the incumbent doesn't care about. Eventually, the disruptive technology gets good enough for mainstream customers—and when it does, it arrives with a cost structure the incumbent literally cannot match without dismantling its own business.

That's the dilemma. It's not a failure of vision. It's a structural trap.

The Casualties: When Input Costs Collapse

The pattern becomes most devastating when the raw cost of a critical input drops by an order of magnitude or more. Let's walk through what actually happened in industries where this occurred.

Minimills vs. Integrated Steel

In the 1960s and 70s, companies like Nucor began using electric arc furnaces to melt scrap steel. The technology was crude. The output was low-quality rebar—the least profitable product in the steel portfolio. Bethlehem Steel and US Steel looked at minimills and saw nothing threatening. Rebar margins were terrible. Let the upstarts have it.

But minimills had a 40–60% cost advantage baked into their architecture. They didn't need massive blast furnaces, coking operations, or the thousands of workers required to run them. As they improved their metallurgy, they moved into structural steel, then sheet steel, then flat-rolled products. Each time they climbed upmarket, the integrated producers retreated to the next "safe" segment.

By the time the integrated mills recognized the existential threat, their cost structures made it impossible to compete. Bethlehem Steel filed for bankruptcy in 2001. Nucor became the largest steel producer in America.

The integrated mills weren't stupid. They were rational. They kept investing in their highest-margin products and their best customers. Christensen would point out that this is exactly what killed them.

The Disk Drive Death Spiral

This was Christensen's original case study and it's almost eerie in its consistency. The disk drive industry went through successive form factor transitions: 14-inch to 8-inch to 5.25-inch to 3.5-inch to 2.5-inch. Each transition followed the same script.

The leading companies in the larger form factor developed the smaller technology first. They had working prototypes before the startups did. But when they showed the smaller drives to their enterprise customers, the response was always the same: "Why would we want less storage capacity? Give us more of what we already have."

So the engineers who championed the new form factor left, joined startups, and sold the smaller drives to new markets that the incumbents didn't serve—minicomputers, then desktops, then laptops. Each time, the smaller form factor eventually met the needs of the larger market, and each time, the incumbent was displaced.

Across every single transition, the leading incumbent almost never became the leader in the next form factor. The same engineers. The same fundamental technology. Completely different outcomes based on organizational structure and incentive alignment.

Kodak and the Zero Marginal Cost Image

Kodak is the example everyone knows, and it's the one most people get wrong. The standard narrative is that Kodak "missed" the digital revolution. They didn't. Kodak invented the digital camera in 1975. They had more digital imaging patents than almost anyone. They understood exactly what was coming.

The problem was structural. Kodak's profit model was built on consumables: film, chemical processing, and photographic paper. Every image someone took required physical materials that Kodak sold at high margins. Digital photography reduced the marginal cost of an image to zero. There was no version of Kodak's business model that survived that transition intact.

Kodak tried everything. They launched digital cameras. They created an online photo service. They invested billions in the transition. But every dollar they put into digital cannibalized their enormously profitable film business. The organization's immune system rejected the new model at every turn. Middle managers optimized for film revenue. Sales teams pushed film products. The quarterly earnings call rewarded film growth.

Kodak filed for bankruptcy in 2012, holding a patent portfolio that proved they had seen the future more clearly than anyone.

Long-Distance Telecom: From Dollars to Dust

In the 1980s, a long-distance phone call cost real money. AT&T, MCI, and Sprint built enormous businesses on per-minute billing. Then the cost of transmitting data over fiber collapsed. VoIP emerged. The cost per minute went from dollars to fractions of a penny.

AT&T, the most iconic technology company of the 20th century, essentially ceased to exist as originally constituted. The brand survived through a series of acquisitions and mergers, but the long-distance telephony business—once worth hundreds of billions—simply evaporated. No amount of operational excellence could sustain a business model built on charging for something that had become nearly free.

The Pattern Christensen Would Recognize in SaaS

If Clayton Christensen were alive today and you showed him the current state of the SaaS industry, he would recognize the pattern immediately. The inputs to building software—developer time, infrastructure costs, the complexity of writing and maintaining code—are dropping by an order of magnitude. AI-assisted development isn't a sustaining innovation that makes existing SaaS products incrementally better. It's a disruptive force that enables entirely new entrants to build comparable products at a fraction of the cost.

He would note that the first wave of AI-native products will look like toys. They'll have fewer features. They'll lack the deep integrations that enterprise customers depend on. They'll feel rough around the edges. And the incumbents will correctly point out all of these deficiencies while continuing to invest in their existing platforms.

He would also note that the incumbents' most dangerous move is their most natural one: bolting AI features onto their existing products. Adding a chatbot here, an AI assistant there, a "copilot" everywhere. This is the classic sustaining innovation response to a disruptive threat. It makes the existing product better along existing dimensions. It satisfies existing customers. And it completely misses the point.

The disruption isn't that existing products will add AI features. It's that new products can be built from the ground up with AI-native architectures at 10–20% of the cost. That cost advantage compounds into pricing power, faster iteration, and the ability to serve markets that current SaaS pricing makes inaccessible.

Christensen didn't just diagnose the disease. He spent the second half of his career working on the cure. And the companies that successfully navigated disruptive transitions share a remarkably consistent approach.

They built a competitor to their own product.

Not a feature team. Not an innovation lab. Not a "skunkworks" project that reports to the same VP who owns the legacy P&L. A genuinely separate entity with its own leadership, its own economics, its own customers, and—critically—its own permission to cannibalize the parent company's revenue.

Netflix is the clearest modern example. In 2007, Netflix was the dominant DVD-by-mail company. The business was growing. Margins were strong. Reed Hastings looked at streaming—which at the time offered a fraction of the DVD catalog in lower quality—and decided to build a separate business that would eventually strangle the DVD operation. Not supplement it. Replace it.

This was not a popular decision. When Netflix split its pricing in 2011 to force the transition, the stock dropped 77%. Customers revolted. Analysts questioned Hastings' judgment. But Netflix understood something that Christensen had been teaching for a decade: you cannot optimize a new business model inside the organizational structure of the old one. The antibodies will kill it.

Apple did the same thing with the iPhone. The iPod was Apple's fastest-growing product. The iPhone would obviously cannibalize iPod sales. Steve Jobs built it anyway, telling his team that if Apple didn't cannibalize the iPod, someone else would.

In both cases, the successful strategy was what software architects call a "strangler pattern"—you build the new system alongside the old one, gradually routing more and more traffic to the new architecture until the old system can be decommissioned. The key insight is that this is not a gradual migration of the existing product. It's the construction of a fundamentally new product that happens to serve the same market.

The Uncomfortable Prescription

If you're a SaaS incumbent reading this, Christensen's framework suggests the following—and you're not going to like most of it:

First, stop thinking of AI as a feature to add to your existing product. Your customers will love the AI features. Your revenue will grow in the short term. And you will be building an incrementally better horse-drawn carriage while someone else designs a Model T. The AI features are sustaining innovations. They will not save you from disruption.

Second, create a separate team—genuinely separate, with separate economics—tasked with building the product that would put you out of business. Give them a blank sheet of paper and the question: "If we were starting this company today, with AI-native development and modern infrastructure, what would we build?" The answer will not look like your current product with AI bolted on. It will be architecturally different, dramatically cheaper to operate, and initially worse in ways your best customers will loudly point out.

Third, let the new product start at the low end of your market. Serve the customers you're currently pricing out. Serve the use cases that are "too small" for your enterprise platform. This is exactly where disruption always begins, and it's exactly the market segment that incumbents always cede because the margins look unattractive compared to enterprise deals.

Fourth, and this is the hardest part: give the new product explicit permission to cannibalize the old one. Not theoretically. Not in a strategy deck. In the compensation structure. In the board presentations. In the quarterly earnings narrative. If the people building the new product are punished when the old product's revenue declines, they will unconsciously protect the old product. The strangler will never strangle.

The Choice That Isn't Really a Choice

Here's the thing Christensen understood better than anyone: the disruption is going to happen whether incumbents participate or not. The only question is whether they disrupt themselves or get disrupted by someone else.

The history is overwhelming. When the cost of a critical input drops by 90%, the companies built on the old cost structure almost never survive in their current form. The ones that do survive are the ones willing to build the thing that kills their existing business—before someone else does.

Netflix did it. Apple did it. Intel tried and failed because they couldn't let go of x86. Kodak tried and failed because they couldn't let go of film. In every case, the technology wasn't the differentiator. The organizational willingness to self-cannibalize was.

Clayton Christensen passed away in January 2020, just before the AI revolution that would make his work more relevant than at any point since he published it. But his framework doesn't need updating. The industries change. The technology changes. The pattern never does.

The SaaS companies that will thrive in five years are the ones making deeply uncomfortable decisions right now. Not adding AI features to yesterday's architecture, but building tomorrow's architecture from scratch—and giving it permission to eat the business that's paying the bills today.

Christensen would have called it the innovator's only real solution: have the courage to disrupt yourself.