Why we need to rethink scale
In 1966, Bruce Henderson, the founder of the Boston Consulting Group, articulated what would become one of the most influential ideas in the history of business strategy: the experience curve. Its origins date back to T. P. Wright’s original 1936 paper, “Factors Affecting the Cost of Airplanes.” Wright discovered a relationship between the cumulative production of a physical good and the costs associated with producing it. The breakthrough was that you could predict your future cost structure in a way competitors couldn’t.
In 1966, BCG did a major study for a semiconductor firm and made a similar discovery. As Martin Reeves describes it, they found “that a company’s unit production costs would fall by typically 20 to 30 percent in real terms for each doubling of “experience,” or accumulated production volume. This had a profound effect on how companies thought about building cost advantage and pricing their output, specifically moving aggressively to grow in the early years of an industry’s formation, so that they could establish a cost advantage which later entrants would be unable to match, in turn yielding a sustainable competitive advantage.”
The Growth / Share Matrix
This led to one of the most famous strategy frameworks ever created, the BCG growth/share matrix. It followed experience curve logic. If you had a large market share in a high-growth industry, you had the potential to win, and win big! If your relative market share was high (compared to other firms), and the growth rate of the market you were participating in was also high, those businesses were “stars.” They would justify investment. If you had a high share, but the market was growing slowly, those were deemed “cows.” As the nickname suggests, cows were there to be milked, but not to command investment. If the market growth rate was low and your share was low, these businesses were “dogs,” and were candidates for divestment. And to this day, nobody ever figured out what to do with the question marks (or problem children), businesses with low shares but high growth rates.
The idea took off like wildfire, for a management framework, anyway. By the 1980s, according to surveys and observations, roughly half of all Fortune 500 companies were using it to allocate resources across their portfolios. Henderson’s insight shaped much of management thinking of the time: the rise of the conglomerate, the dominance of the Fortune 500, the global disposal of entire sectors (to the point at which American companies could no longer compete in consumer electronics, and a whole lot of merger & acquisition activity as a market for market share emerged in its own right. While later empirical research challenged the validity of the matrix (my personal favorite is a piece called “The Product Portfolio and Man’s Best Friend”), it had an outsize impact on executive thinking for decades, particularly emphasizing the importance of scale.
In a dematerializing economy, conventional views of scale aren’t relevant
As more and more of the value of a modern large firm consists of intangible assets, the taken-for-granted value of scale is questionable. A revealing metric is revenue per employee. In the industrial economy, a high-performing company might generate $200,000 to $500,000 per employee annually. That number was treated as a kind of ceiling—you could push it higher with efficiency programs and automation, but the basic ratio of output to headcount was relatively stable.
Digital businesses are rewriting the rules. Cursor, the AI-assisted software development tool, reached $500 million in annual recurring revenue with fewer than 50 employees—roughly $10 million per person. Midjourney, which has never taken a dollar of external funding, crossed $200 million in revenue with approximately 40 people. A startup called Base44 sold for $80 million six months after its founding, built entirely by one person.
The experience curve assumed that accumulated volume drove down costs. But if your primary input is intelligence rather than labor, and intelligence is available on demand at near-zero marginal cost, the curve collapses. A small team with the right AI can match the output of an organization ten times its size, at a fraction of the overhead.
What Scale Actually Bought
To understand why this matters strategically, it helps to be precise about what scale was actually purchasing in the old model. Scale bought four things: production efficiency, through the experience curve; market power, through the ability to invest in distribution, marketing, and sales infrastructure that smaller competitors couldn’t afford; talent aggregation, through the ability to attract specialized people by offering stability, resources, and career development; and organizational resilience, through redundancy and the ability to absorb shocks that would destroy a smaller organization.
AI is substituting for all four, to varying degrees. Production efficiency no longer requires accumulated volume when a single developer with AI tools can write code, design interfaces, and manage infrastructure that would have required twenty people five years ago. Market power through distribution is being disrupted by AI-driven content, community-based growth, and platforms that allow small producers to reach global audiences without large sales organizations. Talent aggregation is becoming less critical when AI agents can perform an expanding range of specialized tasks. And organizational resilience, while still a genuine advantage of large organizations, matters less when smaller organizations can operate with lower fixed costs and therefore survive disruptions that would force layoffs and restructuring at larger rivals.
The new barriers to entry?
The evidence suggests three durable new “moats” that take over from scale. The first is proprietary data. A company that has accumulated unique, hard-to-replicate datasets has an advantage that AI amplifies. The more powerful the AI tools, the more valuable the proprietary data becomes. This is why the most defensible AI businesses tend to be deeply vertical: they have access to data that generalist competitors cannot replicate. Companies as varied as Netflix, Spotify and John Deere keep their strong positions because only they have access to crucial information about their customers.
The second is trust and relationships. In a world where AI can generate convincing content, proposals, and analysis at scale, the scarce resource becomes authentic human connection and trust. Customers of professional service firms, healthcare providers, and any business where the buyer is taking a significant personal or financial risk will continue to value relationships that go beyond what an AI-mediated interaction can provide. Edward S. Jones, Delta Airlines and Zurich Insurance have leaned into human interaction as a differentiator.
The third is what we might call ecosystem position—the ability to sit at the center of a network of complementary actors whose collective value exceeds what any single participant could create alone. Platform businesses, community-driven products, and companies that serve as connective tissue in an industry retain meaningful advantages that scale amplifies rather than creates. Apple, for instance, has such a powerful ecosystem position that Google pays it around $20 billion to be the default search application on the phone.
The New Disruption
For strategists, the implication is uncomfortable but clear. In a world of what Scott Anthony calls “Epic Disruptions,” there are two mechanisms that lead to scale based advantages simply evaporating. The first is when something that used to be complex or incredibly difficult becomes easy. The second is when something that used to be expensive or inaccessible becomes affordable. Combine these two factors, and you have the modern manifestation of Clayton Christensen’s innovator’s dilemma, but today in increasingly digital form.
The assumption that scale will save you is dangerous.