Strong Ecosystems Grow from Parts That Know How to Work Together
An InfoMoney article revisited a piece of advice Steve Jobs once gave Larry Page, at a time when Google had already become a sprawling company. Jobs reportedly urged Page to cut aggressively, narrow the field and concentrate on a handful of products that truly mattered. The reasoning was familiar: the more things an organization tries to do at once, the more likely it is to become merely competent at many of them instead of exceptional at a few. The article frames the later creation of Alphabet almost as evidence that Page did not follow that advice. But the opposite may be closer to what actually happened.
Page seems to have absorbed the principle, applied part of it inside Google, and then changed its scale. Instead of forcing an enormous organization to choose only a few directions, he created an architecture in which different units could pursue their own directions, define their own priorities and, above all, maintain their own focus. Not less focus, then. Distributed focus.
It brings to mind a well-known chef and restaurant owner in São Paulo who argued that a good restaurant should not have an enormous menu. Ten or twelve dishes, carefully prepared and consistently executed, could be enough. I also remember another restaurant, much more popular and nearly always crowded, with tables and benches made from solid wooden planks. Its entire menu, also small, was written on a huge board fixed to the wall where everyone could see it. Few choices, but the restraint seemed to be part of the place’s strength.
A short menu is not necessarily a sign of limited creativity. It is a decision about where attention should go. Fewer ingredients, fewer inventories, fewer procedures, fewer sources of error, and far more repetition. Repetition, when it is not merely mechanical, becomes refinement. A dish prepared hundreds of times is understood differently from one that appears occasionally among dozens of alternatives.
Jobs’s logic seems close to this. Apple spent long periods concentrating on relatively few product families, developed through tight integration among hardware, software, design, services and user experience. That did not prevent deep change within those products. Focus does not necessarily mean immobility.
The Porsche 911 shows this from another angle. It has remained in continuous production since the 1960s and, although instantly recognizable, almost everything about it has changed over time. Engines, suspension, electronics, materials, safety, aerodynamics and internal systems have been repeatedly updated. Countless special versions have appeared, both from Porsche and from companies outside it. There is continuity without freezing.
The Volkswagen Beetle followed a different path. Production ended, but the platform remained culturally and technically alive. Parts, accessories, restorations, conversions and custom projects allowed the architecture to continue evolving without depending directly on Volkswagen. One model kept changing inside its manufacturer; the other kept changing outside it as well. In both cases, something sufficiently stable remained to allow transformation without forcing everything to begin again from zero.
That may be close to what Page understood when Alphabet was created. Google was no longer simply a product, nor even a company with a few variations of the same activity. Search, advertising, artificial intelligence, autonomous mobility, computing, health and other fields were beginning to require different operating logics. Keeping all of them under the same center could produce exactly the problem Jobs had warned about: dispersion. The answer, however, did not have to be simply eliminating possibilities. Some of them could be separated.
This is where modular structure appears. The whole grows, but not as one undifferentiated mass. It is divided into units capable of concentrating attention internally, while a higher level preserves direction, resource allocation and coherence. Jobs can still make sense inside the modules. Page adds the possibility that several focused modules can exist at the same time.
Military hierarchy offers an old analogy. A commander does not direct thousands of soldiers individually. There are units, subunits and intermediate levels of command. Roman legions already operated through layered organization, distributing people among smaller groups under different levels of leadership, including centurions. The point was not hierarchy for its own sake. A mass without structure is difficult to coordinate.
But the reverse problem also exists. If every small group required multiple layers of command, the system would become too heavy. If one officer were responsible for too many people, control would collapse. If responsible for too few, unnecessary bureaucracy might emerge. Organization depends on proportion.
Companies live under a similar tension. Presidents, directors, managers, coordinators, teams and other layers vary widely from one organization to another, but one function keeps returning: preventing every decision, every signal and every problem from reaching the same center directly. Each level absorbs part of the complexity.
Modularizing, then, is not the same as creating departments. A company can have dozens of departments and still be badly organized. A module needs to know what it does, how far it can decide, how it relates to the others and what must be shared. Separation without connection produces silos. Connection without enough separation produces constant interference.
The same pattern appears elsewhere. McDonald’s preserves a recognizable global identity without offering exactly the same menu in every country. Some products act as a common core while others respond to local cultures and markets. The system remains itself without demanding total uniformity.
Get CognycopIA Nexhowl - Estúdio Laboratório’s stories in your inbox
Join Medium for free to get updates from this writer.
Something similar happened in the evolution of media. Phonographs and gramophones gave way to records, tapes, CDs, digital files and eventually streaming. This was not merely a sequence of better objects. The organization of the system itself changed. At one stage, music was strongly tied to physical media; later it moved through files; eventually much of the experience shifted toward platforms, catalogs, networks and continuous access. A new layer began organizing what had previously depended on another structure.
The evolution of companies can be read in the same way. The artisan could concentrate creation, production, quality control, customer relations and often sales in a single person. As production expanded, those functions separated. Larger workshops appeared, then factories, departments, corporations, holdings, supplier networks and business ecosystems. Growth did not simply mean putting more people around the same activity. The pattern of relationships had to change.
Drucker helps explain why this matters for innovation. He did not treat innovation only as the result of a brilliant idea appearing unexpectedly. He emphasized systematic attention to opportunities: unexpected events, incongruities, process needs, changes in markets and industries, demographic shifts, changes in perception and new knowledge. The field of observation is broad, but that does not mean doing everything.
An organization can perceive many opportunities and choose only a few. It can also keep some of them in different spaces until it becomes clearer which deserve to grow. In that sense, Drucker opens the field, Jobs forces selection, and Page adds an architecture capable of sustaining more than one selection at the same time.
There is a larger question underneath all of this: what should remain in the core, and what can exist at the edge? A mature activity has to work today. Another may be testing something that could work tomorrow. Forcing both to obey the same rhythms, metrics and tolerance for risk may damage both.
Modular structures can distribute not only tasks, but uncertainty. Some modules preserve stability; others experiment. Some refine; others search. Projects can begin small, grow, be absorbed into the core, merge with others or simply disappear. Change can remain local without requiring the entire system to be rebuilt around every attempt.
But modularity is not an automatic solution. Every new unit adds interfaces. Meetings appear, information has to move, resources become contested, dependencies grow, metrics multiply and coordination becomes more demanding. The more a system is divided, the more carefully it has to reconnect itself. At some point, additional modules can create more complexity than they were supposed to solve.
Perhaps that is why some things should not remain confined to specific modules. Transparency, communication, flexibility, openness to suggestions, quality and integrity need to move across levels. This does not mean that everyone should know everything, or that every decision should be collective. It means people need enough understanding of the system they participate in to avoid working blindly.
Communication cannot move only downward either. The people executing an activity are often the first to notice a failure, a waste, an unmet customer need or an opportunity for improvement. An organization that distributes responsibility but does not allow information to move upward uses only part of the intelligence it already contains.
That brings structure closer to participation. If people help identify problems, improve processes, raise productivity and create value, there is a coherent argument for part of that value to return to them. Profit sharing or broader participation in results need not remain concentrated at executive levels. Part of it could even take different forms — cash, retirement support, education, healthcare, professional development, time off or other benefits — provided that it does not replace fair compensation or become a single imposed solution.
There is a structural coherence in that. Distribute functions without distributing information and the system becomes blind. Distribute responsibility without autonomy and it generates frustration. Distribute targets without any meaningful participation in results and an asymmetry begins to appear.
There is also a dimension often treated as secondary: the quality of life of the people doing the work. An organization can appear highly efficient while transferring part of its apparent efficiency into human exhaustion. If productivity depends permanently on overload, the structure may be hiding a weakness behind its indicators. Sustainable systems also have to preserve the people who keep them operating.
None of this means that every company should imitate Alphabet or divide itself into autonomous units. Smaller organizations may work better with far more integrated structures and direct communication. Modularizing without necessity can simply manufacture bureaucracy. The issue becomes relevant when complexity, size, variety or speed begin to exceed what a single structure can handle well.
That may be why the apparent contrast between Jobs and Page is less interesting than it first seems. Jobs identified the danger of dispersion. Page appears to have accepted that diagnosis while finding another response to it. Instead of merely reducing the number of possible paths, he created distinct spaces in which different paths would no longer have to compete continuously for the same center.
A restaurant still has to know how many dishes it can execute well. The 911 has to remain recognizable while changing. The Beetle shows how an architecture can outlive the system that originally produced it. A legion has to turn thousands of individuals into units that can be coordinated. A company has to allow its parts to function without forgetting that they belong to the same whole.
Strong ecosystems may not be those in which everything is centralized, nor those in which every part does whatever it wants. They may be those in which there is enough difference for specialization to emerge, and enough connection for those differences to still compose something larger.
Page may not have stopped following Jobs. He may simply have understood that, once a system grows, focus has to learn how to grow with it.
Modular Structure
Strong Ecosystems Grow from Parts That Know How to Work TogetherPortuguese version available on Substack: https://nexhowledumampo.substack.com/p/estrutura-modular-modular-structure?r=5tj4nj&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true
business ecosystems, management
