Key Takeaways
Treat the network as your real audience, not a channel you control
Customers can amplify experience. In 2007, Starbucks CEO Howard Schultz wrote an internal memo warning that streamlined stores had weakened the coffeehouse experience. Once it leaked, customers, employees, journalists, and analysts debated the company’s choices worldwide. Starbucks discovered that its carefully designed website offered no quick way to join that conversation. Similar cascades followed a customer’s complaint about Dell and musician Dave Carroll’s video about United Airlines damaging his guitar. A 2009 Nielsen study cited in the book found that 90% of customers trusted other customers’ recommendations more than any form of advertising.
Participation beats message control. Networks can turn a single customer’s story into a public verdict. Amazon accepted negative product reviews because helping buyers make informed choices could earn more trust than suppressing criticism. Companies cannot control every conversation, but they can listen, respond, and make the underlying experience worth defending.
The important shift is from treating communication as distribution to treating it as a relationship. A company’s public story is increasingly co-authored by customers and employees, so credibility depends less on polished messaging than on whether daily operations withstand scrutiny. Yet viral anecdotes are memorable precisely because they are unusual; not every complaint becomes a global crisis. The practical lesson is not to panic over every post, but to build reliable ways to notice recurring problems and respond before frustration becomes a shared narrative.
Sell the job your product performs, not the object itself
Value lives in use. The book describes an economy shifting from mass production toward services. It reports that services accounted for about 80% of the US economy and argues that information technology, saturated product markets, and urban growth were helping drive the shift. A service is not simply a standardized process: it is an experience shaped by context and by the customer’s participation.
Products provision services. The book calls a product a “service avatar,” meaning a physical expression of the service it enables. A car embodies transportation, while an iPhone provides access to communication, media, and apps. Amazon’s Kindle illustrates the point especially well: its usefulness includes a store, reviews, reading, and synchronization across devices. Firms that focus only on the object risk overlooking what customers actually want done.
This perspective resembles the distinction between selling a tool and solving a user’s problem. It helps explain why a product’s surrounding support, access, and convenience can matter as much as its specifications. It also changes how firms define competitors: a transport service competes with other ways of getting somewhere, not merely with similar vehicles. The caution is that physical products remain essential in many industries. Treating them as service components clarifies their purpose, but does not erase manufacturing, reliability, or ownership as sources of value.
Move decision power to the customer-facing edge
Services meet the unexpected. Factory systems work best when inputs are predictable and outputs can be standardized. Customer needs rarely cooperate. People arrive at inconvenient times, bring problems that do not fit a menu, and expect help in context. The book argues that service organizations should optimize the point of interaction, not only the production line behind it.
Give the front line authority. Vanguard offers a contrast to the usual call-center maze. Customers reaching the investment company speak to a person immediately. If that representative needs an expert, the expert joins the call rather than leaving the customer to restart the story. The representative owns the problem and can draw on the company’s wider knowledge. This requires investing in employees, information access, and judgment, rather than measuring success only by short calls or low support costs.
The design principle is to place discretion where information is freshest. Front-line workers often see the customer’s full situation before a distant manager does, so authority can reduce both delay and repeated handoffs. This is not an argument for eliminating standards: safety, privacy, and legal obligations need reliable controls. The useful distinction is between rules that protect the customer and rules that merely transfer inconvenience to them. A local decision is powerful when the worker has enough context, training, and support to make it responsibly.
Let customer purpose, not quarterly profit, define success
Purpose directs learning. A company’s purpose, in this account, is the job it does for customers. Profit matters, but it is an outcome of creating value, not proof by itself that value is being created. Fred Reichheld’s distinction between good and bad profits helps make the point: repeat business earned through useful service differs from earnings extracted through nuisance fees or customers who feel trapped.
Feedback tests the promise. Net Promoter Score, or NPS, asks customers how likely they are to recommend a company, then sorts answers into promoters, passives, and detractors. Enterprise Rent-A-Car built its own version around customer satisfaction, surveying at least 25 customers per branch each month and following up on poor experiences. The book reports that dissatisfied customers fell from 12% to 5% after the system was introduced. Feedback matters when it leads to concrete improvements, not simply a score on a dashboard.
The useful distinction is between a measure and the reality it attempts to represent. Recommendation scores can expose relationship problems that revenue figures miss, and follow-up gives a number a route back to action. But no single metric captures customer value for every business, and incentives tied too tightly to one score can invite gaming or neglect of unmeasured needs. The stronger practice is to pair a simple signal with the customer’s explanation and evidence about what happened, then give teams the authority to address root causes.
Build small autonomous pods on shared platforms
Small units preserve ownership. A pod is a small, autonomous team responsible for delivering a complete outcome customers value. Unlike a department devoted to one function, a pod can bring different specialties together around a service. The book uses Herbert Simon’s watchmaker story to explain modularity: one watchmaker lost little work when interrupted because he built watches from subassemblies, while the other had to restart whenever his assembly fell apart.
Platforms connect the pieces. A platform is shared infrastructure, standards, information, and support that help independent teams work together. Whole Foods organized stores into self-managed teams, gave them access to performance data, and connected team leaders into larger coordinating groups. Netflix’s streaming operation, as described in the book, used hundreds of independently managed software services. The pattern pairs local freedom with common interfaces, allowing units to adapt without requiring every change to be coordinated centrally.
Modularity is a way to limit the damage caused by change: a team can adjust its work without forcing the whole organization to stop. Shared platforms provide the coordination that autonomy alone cannot supply. This resembles resilient engineering, where systems are designed to isolate faults rather than assume nothing will fail. The tradeoff is real. Separate teams can duplicate effort, disagree about priorities, or optimize their own results at the organization’s expense. Clear purpose, shared measures, and dependable interfaces are what make independence productive rather than merely fragmented.
Run many small strategic experiments before making one big bet
Plans cannot predict every turn. In an uncertain market, the book favors emergent strategy: a strategy that develops through experiments across the organization, rather than being fully specified by senior leaders in advance. Small trials create evidence and options; promising ideas receive more resources, while weak ones are stopped.
Make experimentation routine. Google’s 70-20-10 approach, as presented in the book, directed most engineering effort toward its core business, some toward employee-chosen projects, and a smaller share toward scaling promising ideas. Google also tested thousands of software changes annually and adopted a fraction of them. Amazon similarly tested features with limited audiences before broader release. Experiments should begin as hypotheses that could be disproved, not as proof exercises designed to validate a decision already made.
A portfolio of experiments treats strategy as disciplined search. It can reveal opportunities that leaders would not generate through forecasting alone, especially when information is distributed across teams and customers. The method also protects a company from tying its future to a single prediction. Still, experimentation is not automatically wise: too many trials can consume attention, and local wins may not scale. Each test needs a learning question, a bounded cost, and a credible path from evidence to a decision. Otherwise, experimentation becomes activity mistaken for discovery.
Give leaders the job of connecting, not controlling
Leaders make reality visible. In a connected company, leadership means clarifying purpose, building trust, and helping people understand what is happening around them. The book calls the pressures that push an organization to adapt “adaptive tensions.” Leaders make these pressures tangible by bringing customer experience and market changes into everyday conversations.
Listen across the system. Vanguard circulated customers’ comments to employees through its Voice of the Client program. GE’s Jack Welch and IBM’s Lou Gerstner both pushed leaders to hear directly from customers and challenge internal routines. Leaders also need to connect groups, surface useful ideas, and model the values they expect others to follow. If a company claims service matters, executives should experience customer support themselves, not leave it to the lowest-status employees.
This account replaces the leader as command center with the leader as a high-quality sensor and interpreter. A senior person can connect information from distant parts of an organization and help teams see how their actions relate to a shared aim. That role is especially valuable when reports hide the texture of customer experience. Yet listening without follow-through is theater. Leaders must also make tradeoffs, resolve conflicts, and protect people who raise unwelcome facts. Connection earns authority when it changes decisions, not merely when it produces more meetings.
Make room for freedom, but set boundaries that protect the whole
Autonomy needs guardrails. The book’s Law of Requisite Variety says a control system needs enough range to handle the variety it faces. A fast-food chain can narrow customer choices and standardize the experience. A hotel or investment firm may need workers who can respond to many unanticipated situations. The right balance depends on the service.
Do not confuse freedom with safety. Too little autonomy can make a new venture inherit the constraints that crippled its parent. Too much can expose the whole company to harm. The book points to AIG’s financial products unit, whose risk-taking created liabilities that threatened the company, and to Yahoo’s Brickhouse innovation studio, which lacked sufficient support and protection from the parent organization. Shared purpose, financial limits, common standards, and clear accountability let teams act without making every decision a company-wide gamble.
The challenge is not choosing between control and freedom, but deciding which decisions belong where. Risk that is reversible and close to customers can often be handled locally; risks that threaten solvency, safety, or trust need organization-wide boundaries. This is a governance problem as much as an organizational-chart problem. The examples also show that autonomy requires a platform: resources, legitimacy, and connections to the parent. A team isolated in the name of freedom may be just as vulnerable as one buried under approvals.
Start connectedness with one real customer-facing experiment
Begin where the customer is. The book outlines several routes toward a connected company: grow organically around small teams, change the organization through committed leadership, create a protected pilot pod, or strengthen networks across existing groups. The common first move is direct contact with real customers and the people serving them.
A pilot must be allowed to learn. O2 created giffgaff as a low-cost mobile service shaped with customers and supported by customer contributions. The book reports that it operated with fewer than 20 staff, roughly 5,000 subscribers per staff member, and a Net Promoter Score of 73%, compared with an industry average of 19%. Social tools alone are not enough: if employees can talk to customers but cannot solve their problems, the company has built a connected prison, not a connected business.
A pilot is valuable when it tests a way of working, not just a product idea. It should face real customers, have enough independence to act, and remain connected to resources that can help it survive. This makes learning practical and exposes institutional barriers that a presentation cannot. The giffgaff example is striking, but its results do not guarantee that the model transfers to every industry. The transferable lesson is narrower and more useful: choose a customer problem, give a small team authority to address it, and learn from what customers actually do.
Analysis
The book’s central argument is that organizational structure is not neutral. It determines what a company can notice, decide, and learn. A hierarchy optimized for repeatable production makes sense when variation is low and demand is stable. When customers can compare alternatives, share experiences, and expect tailored service, the same structure can turn ordinary exceptions into delays, handoffs, and resentment. The proposed alternative is not the abolition of management. It is a different division of labor: autonomous customer-facing teams handle local variation, while platforms provide shared standards, resources, and memory.
The book is strongest when it connects human behavior to system design. A customer-service representative cannot reliably act with empathy if every useful remedy requires permission. A team cannot adapt if its incentives reward only its own department’s results. In this sense, culture is not a poster or a speech; it is the pattern of decisions the system makes easy. The examples of Vanguard, Whole Foods, and Amazon show how authority, information, and customer feedback can be arranged to support faster learning.
Its biological and network metaphors are powerful but should not be mistaken for literal prescriptions. Companies are not ecosystems without accountability. A market can reward harmful behavior, and local autonomy can magnify financial, ethical, or safety risks. Nor does a customer score fully represent value: customers may disagree, remain silent, or prioritize needs a survey misses. A connected company therefore needs governance, not just connectivity, and judgment, not just data.
The most durable insight is a design test: place flexibility where uncertainty is high, and stability where coordination and reliability matter most. Then shorten the distance between action and feedback. This avoids both extremes, the brittle machine that cannot respond and the chaotic network that cannot coordinate. The book’s ideal organization is neither flat nor rule-free. It is a learning system with clear purpose, distributed initiative, and shared infrastructure that keeps local action aligned with customer value.
Review Summary
The Connected Company receives high praise for its insights on adapting organizations to the modern, connected world. Readers appreciate the accessible writing style, practical tips, and engaging examples. The book's focus on customer-centric approaches, decentralized decision-making, and fostering a culture of learning resonates with many. Some reviewers note its relevance for large corporations seeking to innovate, while others find it valuable for entrepreneurs. A few critics mention that the book may oversimplify complex topics or rely too heavily on anecdotal evidence.
People Also Read
Glossary
Service-dominant logic
Value created through serviceA way of understanding business in which products are tools for delivering services, rather than ends in themselves. Customers participate in creating value through use and interaction, while knowledge and the ability to respond to context become central competitive strengths. It contrasts with a production-centered view that treats customer service as a cost added after manufacturing.
service avatar
A product expressing a serviceA physical product understood as the material expression of one or more services. A car, for example, gives tangible form to transportation, while a phone provides access to communication and digital services. The term shifts attention from the object alone to the capabilities, knowledge, and customer experience that the object makes available.
pod
Autonomous customer-value teamA small, self-directed organizational unit with authority to deliver an outcome valued by customers. A pod brings together the capabilities needed to act with limited dependence on other units. It can change its methods locally, learn from customer interaction, and remain accountable to a shared purpose and agreed boundaries.
podular organization
Organization built from podsAn organizational design that combines autonomous teams into a larger whole. Instead of grouping all workers by specialty, it forms small units that can deliver complete services and stay close to customers. Shared platforms and standards connect the units, allowing the organization to gain flexibility without becoming entirely disconnected.
platform
Shared support for a communityA support structure that helps a group coordinate and do more together. In a company, a platform can include common technology, information, standards, training, and governance. It enables teams or partners to operate with some independence while sharing infrastructure and rules that reduce friction and protect the wider system.
Learning field
Space for shared learningA designated setting or period in which people can reflect on experience, exchange knowledge, and improve their work. Teams, apprenticeships, and structured employee workshops can serve as learning fields. They help turn individual, often tacit know-how into insights that others can discuss, use, and build into shared practice.
successive approximation
Improvement through repeated feedbackA learning process in which people act, observe the result, reflect, and adjust their next attempt. The process gradually improves performance without requiring a complete plan in advance. In organizations, it underlies iterative product development and experimentation, because each cycle produces feedback that informs the next version.
Law of Requisite Variety
Response range must match variationA principle associated with cybernetics that says a control system needs enough possible responses to handle the variety it faces. An organization can either reduce environmental variety through standardization or increase its own capacity to respond. When customer needs are diverse, rigid systems may be unable to absorb them.
Pace layers
Layers changing at different speedsA way to organize a system by how quickly its parts change. In a company, purpose and core infrastructure may remain stable, while customer interactions, products, or experiments change rapidly. Designing for different rates of change means keeping volatile activities flexible and giving slower-changing foundations enough stability to support them.
emergent strategy
Strategy formed through experimentsA strategy that develops from actions and learning across an organization, rather than being fully planned by senior leaders at the outset. Teams test possibilities, successful approaches attract further resources, and weak ones are stopped. It complements deliberate planning by preserving options in uncertain conditions.
Net Promoter Score
Recommendation-based customer measureA customer-feedback measure based on asking how likely a respondent is to recommend a company or service. Responses are grouped as promoters, passives, or detractors, and the score is calculated by subtracting the percentage of detractors from the percentage of promoters. The book emphasizes pairing the score with follow-up questions and action.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.