Most companies think about innovation in the wrong place. They focus on the product — how to make it better, faster, cheaper. They invest in process improvement — how to deliver it more efficiently. Both matter. Neither is where the most significant competitive advantage lives.

The businesses that create durable competitive advantage typically do it by rethinking something more fundamental: the model through which value is created, delivered, and captured. That’s business model innovation. And it’s where the leverage is.

Netflix didn’t outcompete Blockbuster by offering better movies. It rethought what kind of business a video rental company was, and then built a fundamentally different economic engine around that answer. AWS didn’t win the cloud computing market by building better servers than IBM. It rethought what computing infrastructure was — from capital expenditure to utility service — and changed the economics of an entire industry in the process.

The challenge is that business model innovation is the hardest kind to see coming, for incumbents and new entrants alike. It’s also the hardest to defend against, because it often changes the rules of the competitive game rather than playing them better.

What Business Model Innovation Actually Is

A business model describes how a company creates value for customers and how it captures a share of that value for itself. Business model innovation is a change to one or more of these elements that fundamentally alters the economic logic of the business.

It’s worth being precise about what this is not. Launching a new product is not business model innovation — it’s product innovation. Reducing your manufacturing costs is not business model innovation — it’s operational improvement. Entering a new geographic market is not business model innovation — it’s market expansion. These are all legitimate and often important. They’re just not what this is.

Business model innovation involves changing something structural: who you’re serving, what you’re offering, how you deliver it, or how you charge for it — in a way that creates value differently or captures it from a different source.

Three types dominate the current period of business model transformation:

Subscription models that shift large upfront costs to predictable, recurring revenue. The economic effect is substantial: companies with usage-based or subscription pricing consistently achieve higher net revenue retention than those on traditional licensing structures — often above 120%, meaning existing customers expand their spending year over year. Salesforce didn’t just build better CRM software than Siebel. It changed the economic relationship between the customer and the product.

Platform models that create value by facilitating interactions between two or more distinct groups rather than producing a product or service directly. Airbnb doesn’t own hotels. Uber doesn’t own cars. The platform model’s power comes from network effects — each additional participant increases value for all other participants, which makes the platform increasingly defensible as it grows.

Freemium models that offer a core capability at no cost to drive adoption, then monetize through premium features, enterprise contracts, or adjacent services. The bet is that the cost of acquiring users through free access is lower than the lifetime value of the subset who convert to paid. Figma, Notion, Slack — the model has proven durable in B2B software, and it’s increasingly being applied outside it.

Where to Look for Business Model Opportunities

The most valuable question in business model innovation isn’t “what model should we use?” It’s “where is value being created that isn’t being captured — and why?”

Three sources are consistently productive.

Customer workarounds. Clayton Christensen identified customer workarounds as one of the highest-signal indicators of unmet need. When customers are assembling solutions from pieces — using spreadsheets as a database substitute, using email as a project management tool, cobbling together multiple services to do something they’d pay to have done simply — they’re revealing that the available solutions don’t adequately serve their actual job.

The workaround is the market speaking. It’s telling you what customers would pay to have solved if someone solved it well. Instagram started as a check-in app called Burbn. The founders noticed users were almost exclusively engaging with one feature: photo sharing. They stripped everything else and rebuilt around what customers were actually doing. The product that emerged sold to Facebook for a billion dollars.

Non-consumption. The other signal Christensen identified as underexplored is who isn’t buying at all. The traditional market framing asks “how do we get more customers like our current customers?” The business model innovation question asks “who would use this if we served them differently?” The answer is often a different price point, a different delivery mechanism, a different configuration that makes the core value accessible to people currently priced out or excluded by complexity.

Constraint inversion. The most powerful form of business model innovation often comes from finding a constraint that everyone in the industry has accepted as fixed and asking whether it’s actually fixed. Every incumbent accepts certain costs, certain distribution approaches, certain customer acquisition channels as given. The innovator who can make one of those constraints variable — or eliminate it — changes the economics in ways incumbents can’t easily replicate.

AWS is the clearest recent example. The cost of computing infrastructure was a fixed capital expenditure that created significant barriers to entry for any business that needed server capacity. AWS made it variable — pay as you use, at any scale, without upfront investment. This didn’t just create a new business for Amazon. It changed the economics of building technology businesses globally.

The Experimentation Problem — and the Assumption Audit That Solves It

The reason more companies don’t do business model innovation isn’t that they can’t see the opportunity. It’s that they can’t figure out how to test it without betting the existing business on the answer.

This is a legitimate operational challenge. Business model experiments are expensive, disruptive to existing operations, and slow to produce interpretable results. A subscription model test requires customers to change how they pay. A platform model test requires building both sides of the market simultaneously. A freemium test requires absorbing the cost of free users before knowing what the conversion rate will be.

The companies that navigate this well share a specific starting move: before designing any experiment, they run what I’d call an assumption audit — a structured examination of the beliefs embedded in the current model that have never been tested because they’ve never needed to be.

The audit has four questions, each of which surfaces a different class of business model assumption.

What do we believe about how customers discover and evaluate us that we’ve never actually tested? Most companies assume their current acquisition channels are adequate because those channels have historically worked. They’ve never asked whether a different model — inbound-led, community-led, product-led — would be more efficient at their current scale or at the scale they want to reach. The assumption persists because it’s never been made explicit.

What do we believe about pricing that reflects history rather than value? Most pricing decisions are made by reference to past pricing and competitive benchmarks, not by examining what customers would actually pay for specific outcomes. Companies running their first willingness-to-pay research frequently discover that the ceiling is significantly higher than their current pricing — or that a fundamentally different pricing structure (usage-based, outcome-based, tiered by value rather than by features) would unlock a different customer relationship entirely.

What constraints in our current model do we treat as fixed that might be variable? The question isn’t whether you can eliminate the constraint — it’s whether you’ve ever examined whether it’s genuinely fixed or simply assumed to be. Distribution costs, customer service requirements, onboarding complexity — each of these is a candidate for constraint inversion if the right question is asked about it.

Who is not currently buying from us, and what would have to be different for them to start? Non-consumption is the most underexplored source of business model opportunity precisely because it requires thinking beyond the current customer base. The answer usually isn’t “we need to acquire these non-buyers with our current model.” It’s “these non-buyers reveal a version of this market that our current model isn’t designed to serve.”

The purpose of the assumption audit isn’t to immediately redesign the business model — it’s to generate a prioritized list of hypotheses worth testing. The hypotheses that survive prioritization (high potential impact, low cost to test, fastest to produce interpretable results) become the experiment portfolio. The companies that do this consistently — that run assumption audits not as a one-time strategic exercise but as part of their regular planning cadence — find business model opportunities before they need them rather than after disruption makes them urgent.

The Parallel Portfolio Approach

The companies that do business model experimentation well treat it as a parallel portfolio rather than a sequential pivot. They don’t try to transform the core business through a single large bet. They build adjacent experiments that are structurally separate from the core but close enough to leverage its assets — brand, customer relationships, technology infrastructure, distribution.

McDonald’s “Accelerating the Arches” strategy is illustrative. The company didn’t attempt to transform its core fast-food model in a single move. It built digital capabilities alongside the existing business — mobile ordering, personalization, loyalty — that compounded over time into a genuinely different economic relationship with customers, without disrupting the core franchise model that generated the cash flow to fund the investment. The experiment was parallel. The integration came later, once the model was proven.

The logic is important. The core business funds the experiments. The experiments, if they work, eventually become the core business. The transition is managed rather than forced. This is a very different approach from the “burning platform” narrative that often accompanies business model transformation discussions — where urgency is manufactured to justify the cost of disrupting the existing model. The parallel portfolio approach requires neither urgency nor disruption. It requires patience, discipline about what constitutes a real result, and organizational tolerance for maintaining two operating models simultaneously for some period of time.

The Incumbent Blindspot

The hardest business model innovations to see, for incumbents, are the ones that initially look inferior.

Christensen’s disruptive innovation theory describes this dynamic precisely. New models often appear to underperform on the metrics the incumbent uses to evaluate their business. They serve customers the incumbent doesn’t care about, at price points the incumbent can’t profitably serve, with capability levels the incumbent’s existing customers would find inadequate. So the incumbent ignores them, or dismisses them, or acquires the wrong parts of them.

The disruption arrives when the new model improves to the point where it’s good enough for customers the incumbent does care about — and by that point, the new entrant has built cost structures and customer relationships that the incumbent can’t quickly replicate.

The diagnostic question for incumbents is: what assumptions are embedded in our current model that we’ve never examined because they’ve always been true? The answer to that question is where the business model vulnerability lives — and, often, where the next opportunity does too.

Global merchandise trade growth is running below 0.2% in 2025. Companies facing revenue pressure under these conditions have two choices: compete harder for the same customers in the same way, or find where value is being created that their current model isn’t positioned to capture. The first is a margin-compression game with no durable winner. The second is where the next category leaders are being built — not by waiting for crisis to force the question, but by asking it early enough that the answer can be tested before it becomes urgent.