Sustaining vs. Disruptive Innovation

In November 2022, OpenAI launched ChatGPT and pushed one of the world’s most profitable companies into a strategic bind.

For more than two decades, Google improved search through a classic sustaining innovation playbook: faster results, better ranking, richer snippets, stronger personalization, and better advertiser tools. Each improvement made the core product more valuable to users, advertisers, and shareholders. The model was elegant: users searched, Google displayed links and ads, advertisers paid for intent, and publishers supplied the content that kept the system useful.

ChatGPT introduced a different interaction model. Instead of sending users to ten blue links, it synthesized an answer. Instead of making the search results page better, it made the results page feel less necessary.

That is a different kind of competitive threat.

Google’s response has been to integrate generative AI directly into search through AI Overviews and related features. Superficially, this looks like adaptation. Strategically, it is more complicated. Google is being pushed to improve the user experience in a way that may reduce the clicks that support its advertising economics.

This is the innovator’s dilemma in real time. The question is not whether Google has technical capability. It does. The harder problem is economic: the model that may best satisfy future user behavior may be structurally worse for the existing profit pool.

That is why the distinction between sustaining innovation and disruptive innovation matters. Misclassifying the threat leads to the wrong strategic response, the wrong valuation assumptions, and often the wrong capital allocation decisions.

A sustaining threat calls for operational excellence. A disruptive threat calls for business model courage.

Sustaining vs. Disruptive Innovation

A sustaining innovation improves an existing product for existing customers along dimensions those customers already value.

A faster chip, a better camera, a more accurate search algorithm, a more efficient aircraft engine, or a more powerful enterprise software suite — these are sustaining innovations. They make the current product better and usually reinforce the existing business model.

For investors, sustaining innovation usually looks like continued reinvestment into the core franchise. It may support pricing power, retention, margin expansion, or market share defense. The analytical questions are familiar: How durable is the moat? How much capital is required? Does the innovation improve revenue, retention, or unit economics?

A disruptive innovation is different. It usually begins as an inferior product by mainstream standards. It enters through the low end of the market or through non-consumption, where customers are overserved, priced out, ignored, or willing to accept a simpler product with lower performance.

At first, the disruptor may look unattractive. Its customers are less profitable. Its margins may be lower. Its product may appear crude. Its market may seem too small to justify a serious response.

That is precisely why it is dangerous.

The disruptor improves. Its technology, cost structure, distribution, or data advantage compounds. Eventually, it crosses the threshold where mainstream customers begin to consider it good enough. By then, the incumbent faces a dilemma: respond aggressively and cannibalize its own economics, or preserve the existing model and watch the entrant capture future growth.

Disruption is a process of value migration.

The key question is “At what rate is the entrant improving, and when does that improvement intersect with the needs of profitable mainstream customers?”

That crossing point is where valuation risk accelerates.

Why Incumbents Usually Win Sustaining Innovation Battles

In sustaining innovation, incumbents usually have the advantage because the game rewards assets they already possess.

They know the customer. They control distribution. They have purchasing power, engineering depth, brand trust, regulatory knowledge, and the capital to fund large improvement cycles. They can bundle, cross-sell, discount, and acquire capabilities when needed.

This is why established companies often become stronger through sustaining innovation. Apple’s iPhone improvements reinforce the premium smartphone ecosystem. Visa and Mastercard’s investments in fraud detection and tokenization reinforce their payment networks. Microsoft’s integration of AI into Office and Azure strengthens products already embedded in enterprise workflows.

In these cases, innovation does not attack the incumbent’s profit formula. It deepens it.

The financial profile is legible: invest in product improvement, maintain or expand revenue per customer, protect retention, and improve lifetime value. Even if margins compress temporarily, the strategic logic is clear: defend the core by making the core better.

That is why the wrong response to a sustaining threat is panic. If a competitor launches a better product within the same market structure, the answer is usually not to abandon the model. It is to execute better.

Disruptive innovation is harder because the correct response often looks irrational under the metrics governing the existing business.

Why Disruptive Innovation Is Hard to Finance Internally

The most important feature of disruptive innovation is not technological novelty. It is incentive incompatibility.

A disruptive business usually has different customers, margins, sales motions, performance metrics, and capital allocation logic. Inside an incumbent, that makes it hard to evaluate.

The new unit may look unattractive compared with the core. It may have lower gross margins, smaller deal sizes, higher uncertainty, or weaker near-term returns. It may target customers the incumbent has deliberately avoided. It may also threaten to cannibalize the most profitable revenue streams.

From the CFO’s perspective, that creates a real problem. Capital is scarce. Management teams are judged on earnings, margins, free cash flow, and return on invested capital. Boards expect discipline. Investors reward near-term execution. Business unit leaders compete for resources.

In that environment, disruptive initiatives often lose internal capital allocation fights.

Not because managers are foolish, but because they are rational.

That is Clayton Christensen’s most uncomfortable insight: the practices that make companies excellent operators can make them poor responders to disruption. Listening to best customers, improving margins, allocating capital to the highest-return projects, and focusing on the most profitable segments are good management disciplines. But they bias the organization toward sustaining innovation and away from disruptive bets.

The incumbent keeps doing what financial logic tells it to do.

Until the financial logic changes.

Steel Minimills: A Capital Allocation Case Study

The American steel industry remains one of the clearest examples of disruptive innovation because the economics are so visible.

Traditional integrated steel mills required large-scale blast furnaces, iron ore supply chains, significant fixed assets, and high utilization. Their best customers bought higher-grade steel products used in automotive, structural, and flat-rolled markets.

Minimills entered with electric arc furnaces that processed scrap metal. Their capital requirements were lower. Their cost structures were different. Their initial quality was worse. They could not compete for the premium products integrated mills cared about.

So they started with rebar.

Rebar was low quality, low margin, and unattractive to integrated mills. When minimills captured that market, incumbents did not experience it as a crisis. They experienced it as portfolio improvement. Exiting rebar allowed them to focus on more profitable products.

That was rational.

Then minimills improved and moved into angle iron. Integrated mills retreated upmarket. Then minimills attacked structural beams. Again, incumbents protected the higher-margin segments. Each retreat made financial sense. In aggregate, it was strategic surrender.

This is the essence of disruption: rational short-term decisions create long-term fragility.

For strategy professionals, the steel case shows why margin analysis alone can mislead. A low-margin segment may be unattractive on a standalone basis but strategically important as the disruptor’s training ground. The entrant uses the low end to build volume, capabilities, process knowledge, and customer relationships.

For investors, the question is not only whether the entrant’s current margin is attractive. It is whether the entrant’s cost structure allows it to serve a market the incumbent cannot defend without damaging its own economics.

That is where value transfer begins.

Google, ChatGPT, and the Cannibalization Problem

Google’s search business remains one of the greatest business models ever created: high intent, global scale, strong advertiser demand, deep data advantages, and extraordinary operating leverage.

The strategic concern is that generative AI may change the unit of value in search.

For years, Google improved search in ways that supported the existing revenue model. Better ranking increased user trust. Better ad targeting improved advertiser ROI. Featured snippets gave users faster answers while preserving the broader results page. The company could improve the product and monetize the improvement through advertising.

Generative AI creates a harder trade-off. A synthesized answer may satisfy the user more directly, but that same directness can reduce the need to click links. If fewer users click, publishers receive less traffic, advertisers may receive fewer visits, and Google must find new ways to monetize an answer-based interface.

This is not merely a product issue. It is a profit pool issue.

OpenAI, Perplexity, Anthropic, and other AI-native companies do not have the same advertising legacy to protect. They can experiment with subscriptions, enterprise licensing, API usage, premium workflows, and answer-native interfaces without defending a search advertising machine.

That is counter-positioning.

An entrant is counter-positioned when its business model is attractive to customers but unattractive for the incumbent to copy because copying it would damage the incumbent’s existing business. The attacker’s freedom comes from not owning the legacy profit pool.

Google can build AI. The problem is whether the best AI search experience is economically equivalent to the best advertising search experience.

For finance professionals, this matters because a company can remain operationally excellent while its multiple compresses. The business may keep growing, yet the market may begin to discount the terminal value of the core profit engine if the basis of competition is shifting.

Disruptive threats often show up first as narrative risk, then margin risk, then revenue risk.

By the time they appear clearly in reported financials, the valuation debate may already have changed.

Kodak and Blockbuster: When the Model Rejects the Future

Kodak is often framed as a story about technological blindness. It is more accurately a story about economic conflict.

Kodak engineers invented an early digital camera in 1975. The company understood digital photography. The issue was that digital threatened the economics of film, processing, and consumables. Film was not merely a product line. It was the profit engine.

Commercializing digital aggressively would have meant accelerating the decline of Kodak’s best business before the replacement economics were ready. That is a hard decision for any management team. It asks leaders to reduce near-term profitability in exchange for uncertain future relevance.

So Kodak delayed. It protected the old model. The decision was understandable. It was also fatal.

Blockbuster faced a similar problem with Netflix.

Netflix’s DVD-by-mail model did not initially attack Blockbuster’s most valuable use case. Blockbuster was built around physical stores, impulse rentals, late fees, and local availability. Netflix served a more patient customer who valued convenience over immediacy.

At first, it looked niche. But Netflix had a structurally different cost base and customer relationship. It did not need the same retail footprint. Later, when streaming became viable, Netflix was better positioned to move.

Blockbuster tried to respond, but its response was constrained by the old model. It could not fully embrace the economics of the new system while still protecting the physical network.

That is the danger of half-measures. They consume capital without creating strategic escape velocity.

How to Diagnose the Threat

1. Is the entrant serving your best customers or your least attractive customers?

If a competitor is winning your best customers with a better product on dimensions they already value, you are probably facing a sustaining threat. The response is direct competition: improve the product, defend accounts, adjust pricing, strengthen distribution, and invest behind the core.

If the entrant is serving customers you ignore, underserve, or consider low value, pay attention. That is often where disruptive models begin.

The risk is that incumbents define market attractiveness using current economics. Disruptors define it using future improvement potential.

2. Does the entrant have a structurally different business model?

A true disruptor is not simply a smaller version of the incumbent. It has a different cost structure, customer acquisition model, margin profile, or value proposition.

Steel minimills were not smaller integrated mills. Netflix was not a smaller Blockbuster. ChatGPT is not a better list of links. These models changed the basis of competition.

If the entrant’s model allows it to serve demand profitably in a way the incumbent cannot match without self-harm, the threat is more serious than current market share suggests.

3. Would responding properly cannibalize your own margins?

This is the most important test.

If the correct response would reduce revenue per customer, lower margins, undermine a legacy channel, weaken an existing partner ecosystem, or accelerate the decline of a profitable product, then the incumbent has a true dilemma.

The problem is not whether management sees the threat. It is whether management can act against the financial architecture of the company.

That is why disruptive innovation is as much a governance problem as a strategy problem.

When Disruption Becomes Material

For investors, the most dangerous moment is not when the disruptor first appears. Early disruption is usually too small to matter in the numbers.

The dangerous moment is when the disruptor’s improvement curve approaches the performance threshold required by mainstream customers.

At that point, the incumbent’s historical financials may still look strong. Revenue may still grow. Margins may remain attractive. Cash flow may be abundant. Management may continue buying back stock and guiding confidently.

But the future profit pool may already be contested.

That creates a gap between backward-looking financial strength and forward-looking strategic vulnerability. Reported numbers often reward the incumbent until surprisingly late in the cycle.

The analytical task is to identify when the entrant moves from edge use cases into core use cases.

In Google’s case, that means watching whether AI-native answer engines remain tools for students, developers, researchers, and exploratory queries — or whether they begin capturing commercial, local, product, travel, and high-intent searches. The latter matters far more to the economic model.

For enterprise software, the question is whether AI agents remain workflow assistants or begin replacing seats, modules, or entire categories.

For financial services, it is whether embedded finance, stablecoins, real-time payments, or AI underwriting remain adjacent innovations or begin attacking fee pools, spreads, and customer ownership.

Disruption matters when it moves from usage to monetization.

The Operator’s Lens: How to Respond

If an incumbent concludes that it is facing a disruptive threat, the response must be structurally serious.

The worst option is the half-measure: enough investment to claim strategic awareness, not enough autonomy to create a real alternative.

A credible response usually requires a separate unit with its own P&L, leadership, incentives, cost structure, customer targets, and permission to compete with the parent. Without that separation, the new model will be pulled back toward the old one. Sales teams will protect legacy quotas. Finance teams will compare immature economics against mature margins. Executives will seek synergies that quietly destroy the new model’s advantage.

The new business must be allowed to look unattractive by the standards of the old business.

That requires board-level alignment and investor communication. Management must explain why near-term cannibalization may be preferable to long-term displacement. It must set different metrics for the new unit: adoption, retention, learning velocity, cost curve improvement, and progress toward mainstream performance thresholds.

If the company cannot do that internally, acquisition may be the better option. But acquisitions only work when the incumbent preserves what made the disruptor dangerous. Buying the entrant and forcing it into legacy processes often neutralizes the capabilities that justified the deal.

For disruptors, the strategic lesson is the opposite: do not move upmarket too quickly.

The low end is not merely a beachhead. It is a capability-building zone. It gives the disruptor room to improve without provoking the full force of the incumbent too early. Moving into premium segments before the product is ready turns a disruption battle into a sustaining innovation battle — and incumbents often win those.

A Practical Decision Framework

If the threat is sustaining, ask:

• Can we match or exceed the competitor’s product improvement?
• Are we investing enough in the core?
• Do we have pricing power or must we absorb margin pressure?
• Can distribution, brand, data, or scale preserve advantage?
• What is the expected return on incremental defensive investment?

If the threat is disruptive, ask:

• Which unattractive customers or non-consumers is the entrant serving?
• Is the entrant’s cost structure fundamentally different from ours?
• What performance gap currently protects us?
• How quickly is that gap closing?
• Would a proper response cannibalize revenue or margins?
• Can we build a separate model, or must we acquire one?
• What signals show the threat is moving from peripheral to core?

The distinction matters because the response differs.

A sustaining threat requires better execution.

A disruptive threat requires willingness to damage the present in order to own the future.

The Real Cost of Misclassification

Most companies do not fail because they ignore innovation. They fail because they misclassify it.

They treat disruptive threats as sustaining threats and respond by improving the existing product for existing customers. That works until the entrant changes the basis of competition. By then, the incumbent’s capabilities, margins, incentives, and investor expectations are tied to the old model.

The tragedy is that every step can look rational.

Integrated steel mills exited low-margin rebar. Kodak protected film. Blockbuster defended stores. Google is now trying to preserve search economics while adapting to answer-based AI. In each case, the short-term financial logic is understandable.

But strategy is not only about optimizing the current model. It is about recognizing when the current model is becoming the constraint.

Sustaining innovation compounds incumbent advantage, while disruptive innovation transfers value across business models.

The income statement may still look healthy. Margins may still look defensible. The incumbent may still appear dominant. But if the entrant is improving faster than expectations are changing, and if the incumbent cannot respond without cannibalizing itself, the strategic risk is already present.

The distinction between sustaining and disruptive innovation determines whether management should defend the core, build a separate business, acquire the attacker, or accept that future value will accrue somewhere else.

Getting that distinction wrong does not just cost market share.

It can cost the profit pool.