The Innovation Advantage Is Not Growth. It Is Knowing What to Remove.

What if the most innovative organizations are not the ones that create the most, but the ones that remove the most intelligently?

That question looks strange at first. Innovation is usually associated with expansion: more startups, more hiring, more experiments, more capital, more tools. Israel’s digital health ecosystem appears to embody this logic. With a population roughly comparable to New Jersey, it has produced more than 6,000 startups. Its hospitals, entrepreneurs, investors, and technology companies operate in a dense network where new ideas can be tested and transferred quickly.

Yet another pattern is visible inside one of the world’s most successful professional services firms. After its headcount grew from about 28,000 to nearly 47,000 in five years, McKinsey planned to eliminate roughly 1,400 roles, even while reporting record revenue and continuing to recruit. Some managers and associate partners were offered as much as nine months of pay to leave.

These facts seem to point in opposite directions. One looks like a culture of relentless creation. The other looks like contraction inside a highly profitable institution. But together they reveal a more important principle: innovation depends less on how much an organization accumulates than on how effectively it refreshes its capacity to act.

The central organizational problem is not growth. It is metabolic health.

Growth Creates Capacity, Then Conceals Its Cost

Organizations often treat growth as evidence that their strategy is working. More employees suggest confidence. More programs suggest ambition. More partnerships suggest reach. More revenue suggests validation. These signals are useful, but they can also disguise a dangerous shift: the institution may be adding resources faster than it is adding useful action.

Imagine a restaurant that keeps expanding its kitchen. It buys more ovens, hires more chefs, adds more managers, and increases its menu from twenty items to two hundred. For a while, this may look like success. But if orders now require more coordination, ingredients spoil before use, and every decision requires approval from several people, the restaurant has not necessarily become more capable. It has become heavier.

The same thing can happen in knowledge organizations. Headcount grows, but so do meetings, review layers, internal processes, and ambiguous responsibilities. The organization may retain talented people while losing the conditions that allow those people to move quickly. Its nominal capacity rises while its usable capacity falls.

This is why a profitable firm can still need to reduce roles. Financial performance is not the same as organizational fitness. A company can produce excellent results today while carrying a structure designed for yesterday’s demand, yesterday’s priorities, or yesterday’s assumptions about how work should be divided.

The distinction is crucial:

Capacity is what an organization could do if all its resources were available and well coordinated.

Velocity is how quickly it can turn judgment into action.

Adaptability is how easily it can redirect capacity and velocity when conditions change.

Growth often improves capacity first. But without periodic renewal, it can damage velocity and adaptability. The organization becomes like a crowded airport: it has many planes, pilots, and destinations, but too few clear runways.

This helps explain why selective departures can be rational even during expansion. An institution may continue hiring in strategic areas while offering exits in roles whose responsibilities have become redundant, overly layered, or misaligned with future work. The apparent contradiction disappears once growth is understood as reallocation, not simple accumulation.

Israel’s Advantage Is Density, Not Magic

Israel’s digital health ecosystem offers a useful counterpoint because its strength is not merely the existence of talented founders. It is the density of interactions among institutions that can test, fund, refine, and distribute new ideas.

A startup developing a smartphone test for kidney disease does not create value simply by producing an app. The app matters when it enters a clinical and operational system. It needs patients, clinicians, data, regulatory knowledge, implementation partners, and a path to trust. The closer those elements are to one another, the shorter the distance between invention and practical learning.

This is why partnerships between organizations such as Sheba and Mayo matter. Sharing data from startup programs does more than create access to information. It creates a feedback loop. An idea can be evaluated in one institutional setting, compared with experience in another, and improved before its assumptions harden into a business model.

The deeper advantage is compression of the learning cycle.

In a sparse ecosystem, a founder may spend months locating a clinical partner, discovering that the data format is unusable, and learning that a workflow cannot accommodate the product. In a dense ecosystem, those failures can happen sooner. The founder encounters reality before too much capital, time, or institutional identity has accumulated around the original idea.

That is not the same as avoiding failure. It is better: the system makes failure cheaper, faster, and more informative.

A culture that encourages innovation is therefore not simply one that celebrates bold ideas. It is one that permits ideas to be exposed to demanding feedback before they become protected interests. The startup ecosystem benefits from many experiments, but also from rapid filtering. Weak concepts do not need to survive indefinitely in order to justify the people and money already invested in them.

This is the connection to organizational renewal. A healthy innovation system continually answers two questions:

1. What should we start because it may create disproportionate value?
2. What should we stop because it is consuming attention without producing enough learning or impact?

Most institutions are much better at answering the first question. They create innovation labs, appoint transformation leaders, and announce new initiatives. They are much worse at answering the second. Once a project has a budget, a team, and a senior sponsor, stopping it can feel like admitting failure. As a result, old commitments accumulate beneath new ambitions.

Israel’s startup density is impressive, but density alone does not guarantee innovation. What matters is the circulation of ideas through a network that can distinguish promise from inertia. The same principle applies inside a large firm: talent becomes an innovation asset only when it can move toward high value work and away from low value work.

The Hidden Skill Is Strategic Subtraction

The most underdeveloped capability in many organizations is not creativity. It is subtraction.

Leaders are trained to announce initiatives, approve investments, and reward visible expansion. They are less often trained to dismantle structures that once made sense. Yet every organization has a limited supply of attention. Every new priority competes with an existing one. If nothing is removed, the organization does not become more innovative. It becomes more crowded.

Consider a hospital innovation center working with startups. It may have dozens of promising projects, each addressing a real problem. But clinicians have limited time, data systems are imperfect, and implementation teams cannot support every pilot. If leadership refuses to end weak experiments, the strongest ideas may receive too little attention to reach patients. The result is not abundance. It is congestion.

Or consider a consulting firm whose headcount expands rapidly during a period of strong demand. It may hire layers of expertise to serve a broad range of clients and problems. Later, demand changes. Some work becomes automated, some capabilities become less differentiated, and some managerial roles no longer correspond to the firm’s most valuable activities. Continuing to carry every layer because the firm is still profitable would preserve the past at the expense of the future.

This suggests a practical model of organizational metabolism:

• Absorption: Can the organization take in new ideas, people, and information without overwhelming its operating system?
• Conversion: Can it turn those inputs into products, decisions, services, or learning?
• Elimination: Can it release projects, roles, habits, and assumptions that no longer justify their cost?
• Redistribution: Can the freed resources move toward areas with greater future value?

Innovation requires all four. Absorption without elimination produces clutter. Elimination without redistribution produces fear and shrinkage. Redistribution without absorption produces a rigid organization that cannot recognize new opportunities. The goal is not perpetual pruning. It is a healthy cycle of intake, testing, removal, and reinvestment.

The question is not whether an organization is growing. The question is whether growth is increasing its ability to learn.

This reframes difficult personnel decisions as well. A departure program can be handled badly, becoming a blunt cost cutting exercise that destroys trust and institutional knowledge. But the underlying logic of selective renewal is not inherently anti people. In some cases, keeping every role indefinitely is less respectful to employees than acknowledging that the organization’s needs have changed. The ethical test is whether leaders can explain the future they are building, apply criteria consistently, and help affected people move toward their next opportunity.

The same standard applies to innovation projects. Stopping a pilot is not necessarily a rejection of the people who built it. It can be a recognition that the evidence does not support further investment. A mature culture separates the dignity of contributors from the survival of every contribution.

Build Organizations That Learn Before They Scale

The most useful lesson for leaders is to reverse the usual sequence. Organizations often scale first and learn later. They hire a large team, build extensive infrastructure, launch broadly, and only then discover whether the underlying model works.

A more resilient approach is to scale the learning system before scaling the activity.

In practice, that means designing small, fast feedback loops. A healthcare startup should test not only whether its technology functions, but whether clinicians will use it, whether patients trust it, whether the workflow supports it, and whether the data can travel across institutions. A professional services firm should test whether a new layer improves client outcomes or merely redistributes information inside the firm.

Leaders can ask four questions before approving growth:

1. What assumption are we testing? If the answer is vague, the initiative is probably seeking permission rather than learning.
2. What evidence would cause us to stop? If nothing could change the decision, the project is protected from reality.
3. What resource becomes constrained if we add this? Every initiative consumes attention, managerial time, or technical capacity, even when its budget appears small.
4. If this works, what will we remove or redesign? Success should not simply be layered onto the existing system.

The fourth question is especially powerful. It forces leaders to imagine innovation as replacement, not decoration. A new digital diagnostic may reduce the need for certain administrative steps. A new analytical tool may change the role of a team that once performed routine research. A partnership may make an internal process unnecessary. If leaders cannot identify what the new capability will displace, they may be creating parallel complexity rather than progress.

Individuals can apply the same logic to their own work. Maintain a personal inventory of recurring commitments and classify each one as a growth asset, a necessary maintenance task, a learning experiment, or a legacy obligation. Review the inventory monthly. If a task is neither producing value nor generating insight, it deserves a serious challenge, regardless of how long it has been performed.

The objective is not to maximize busyness. It is to increase the ratio of meaningful decisions to organizational effort.

Key Takeaways

• Measure renewal, not just expansion. Track what the organization has stopped, simplified, or reassigned alongside revenue, hiring, and new initiatives.
• Design for fast disconfirmation. Every experiment should state its key assumption, its evidence threshold, and the conditions that would justify stopping.
• Protect density of interaction. Put founders, users, technical experts, operators, and decision makers close enough to one another that feedback arrives before enthusiasm becomes bureaucracy.
• Treat talent as mobile capital. When priorities change, help capable people move toward higher value work instead of preserving roles solely because they already exist.
• Ask what success will replace. A new capability that adds work without removing friction may be expansion disguised as innovation.

The contrast between a compact country producing thousands of startups and a large consulting firm offering generous exits during continued recruitment is not really a contrast between innovation and decline. It is a contrast between two ways of managing organizational energy.

One system gains strength from dense contact, rapid testing, and a willingness to let weak ideas disappear. The other reveals what happens when successful growth eventually creates more structure than the next stage can use. Both point to the same conclusion: the future belongs to organizations that can create and retire with equal intelligence.

We tend to imagine innovation as a spark, a breakthrough, or a new invention. In practice, it is also an act of clearing. The new cannot travel quickly through a system packed with old commitments. Sometimes the most innovative decision is not to add another team, tool, or initiative. It is to remove the obstruction that prevents the best existing idea from moving.