The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses

Chapter 9 advocates for the use of small batches in production. For example, the book Lean Thinking by James Womack and Daniel Hones discusses the authors putting newsletters into envelopes with their children. The children wanted to complete the envelopes step-by-step: fold the envelopes, attach the seal, and put on the stamps. But the adults wanted to create each envelope separately and individually until completion. Counterintuitively, this small-batch method, called “single-piece flow, worked better because of the power of lean manufacturing (184). Single-piece flow can identify issues early in the production process, like if the envelopes and the letters were different sizes and thus didn’t fit. Single-piece flow also accounts for the extra time required to organize the products, such as sorting the envelopes, and removes this process by completing each task individually.

Single-piece flow is an essential part of lean manufacturing history. When Toyota realized that they couldn’t compete with huge, high-tech American factories after World War II, the company purchased smaller, adaptable, general-purpose machines. The machines were adjusted at a rapid pace to make different parts for different steps in the manufacturing. This quick turnover was called SMED (Single-Minute Exchange of Die), developed by Shigeo Shingo. The concept enabled small batch sizes to quickly change their configuration for the part required at each specific step in the process. The smaller batch size of production allowed Toyota to produce a greater variety of automobiles. Quality problems could be identified and solved sooner. Small batches could also allow Toyota to serve smaller markets by creating products unique to the diverse populations of customers. Toyota thus discovered that small batches increase the agility and speed of production.

Ries talks about the small batch product he implemented at IMVU. Every engineer worked together on one feature of the product at the same time. Once that feature was completed, it was released to a small group of customers to gather immediate feedback. IMVU made around “fifty changes to its product (on average) every single day” using this approach” (189). Automated tests were implemented on each feature to catch defects early in the production process. These automatic protections served as an “immune system” for product development (190). The process goes as follows: the defective change is removed, the team is notified of the problem, the team is prevented from introducing any changes to the product, the root of the problem is found and fixed. This process, called “continuous deployment” (190-91), speeds up the Build-Measure-Learn feedback loop, thus ensuring success and sustainability.

Another example of small batch manufacturing is found in education. Instead of following a modern and orthodox, mass production style of education, startups, like School of One, use small batches to teach students. For example, students are given daily “playlists of their learning tasks that are matched to their learning needs” (195). Videos, personal tutoring sessions, and group activities make up the lesson plan, and assessments are built into each activity. These assessments track student success as a set of data that can be evaluated by educators. Following the Build-Measure-Learn feedback loop, teachers can adjust their curricula in small batches to better enable student success in the classroom.

Chapter 9 calls the allure of building with mass production methods the “large batch death spiral” (198). The batches that grow over time eventually force employees to complete additional work to keep up the production. When mistakes occur, the entire production needs to shut down to fix these interruptions. By trying to minimize overhead, the company continues to follow the large batch method without making the necessary fixes to the individual problems in the product. However, with small batches, a company can implement single-piece flow throughout the supply chain. Small batches should “pull” production processes in unison. In car repair, when a bumper is needed, the new one creates a gap in the inventory. This should automatically signal the storage warehouse to supply a new bumper to the dealership, which consequently signals the manufacturer to create a new bumper. Should there be a problem in the bumper on the production line, the manufacturer sends a signal to the product developers who can find the root of the issue. This is called “work-in-progress inventory,” a concept designed by Toyota for lean manufacturing. In the Lean Startup method, the “production development process […] is responding to pull requests in the form of experiments that need to be run” (201).

Chapter 9 ends on the virtues of the Toyota Production System. Not only is the company using the “most advanced system of management in the world, but even more impressive is the fact that Toyota has built the most advanced learning organization in the industry” (204). Toyota uses validated learning, small-batch production, and lean manufacturing techniques to harness the creativity of its employees for constant growth and innovation at a rapid and efficient pace.

Chapter 10 begins by describing two different companies—a collectibles marketplace and a database software enterprise—that both have the same problem: “neither company was growing,” despite using the same engine of growth (207). The engine of growth is “the mechanism that startups use to achieve sustainable growth,” sustainability meaning new customers added through existing customers (207). Customers drive sustainable growth in four ways: word of mouth, side effect of product usage, funded advertising, and repeated purchase or use (208). These sources of sustainability generate the feedback loops called engines of growth. Engines of growth provide startups with metrics that direct the company’s most focused efforts and resources.

As discussed earlier, there are three engines of growth: sticky, viral, and paid. The sticky rate of growth tells entrepreneurs that if the “rate of new customer acquisition exceeds, churn rate,” then the product will grow (211). The churn rate is the percentage of customers who disengage from the product. The former and latter companies tracked activation rates and revenue per customer but failed to track new customer retention and duration of the product’s usage per customer. It’s crucial for a company that provides services, subscriptions, platforms, and applications to have a high customer retention rate. That is, these types of companies need a sticky engine of growth to measure success.

The viral engine of growth applies to social media and other products where consumers do the bulk of the company’s marketing. This is different than word-of-mouth growth because the customer-to-customer transmission is “a necessary consequence of normal product use,” not a supplement (212). Ries uses the example of Hotmail to describe a viral engine of growth. Hotmail originally offered free email accounts to customers, but their growth was sluggish. Instead of spending money on expensive marketing, the company simply added a hyperlinked message to the bottom of each of their emails: “P.S. Get your free e-mail at Hotmail” (213). After a few months the company gained 1 million customers. Hotmail used the viral loop, a quantifiable feedback loop that measures growth. The speed of the viral loop is determined by the viral coefficient: the measurement of “how many customers will use a product as a consequence of each new customer signing up” (213). The higher the viral coefficient, the quicker the product will disseminate among customers. Increasing the viral coefficient is key to measuring the viral engine of growth. It’s important to note that monetary growth doesn’t directly correlate to viral growth; “it is useful only as an indicator that customers value the product enough to pay for it” (215). What viral engines of growth determine primarily are the number of customers investing time and attention to the product, which makes it more valuable to advertisers.

The paid engine of growth measures the amount of money it costs to gain a new customer. Companies can increase this rate of growth by either increasing the profit from each customer or lowering the cost of signing up new customers (216). The paid engine of growth is also powered by a feedback loop, called the lifetime value (LTV): the amount of money a customer spends on a product throughout their lifetime. This engine of growth applies especially to essential products, like food, apparel, home appliances, automobiles, and technology. The revenue from this growth can be invested through the purchase of advertising. If the LTV of a product is greater than the cost of advertising, then the product can grow.

More than one of these engines of growth can operate in a company at a given time. As startups grow and their products and markets diversify, they will need to engage different engines of growth. However, startups should focus on one engine of growth at a time in their beginning stages. With a strong leap-of-faith hypothesis and a strategy for validated learning, a successful engine of growth can be activated.

The end of Chapter 10 highlights Marc Andreesen’s theory of product/market fit. “Product/market fit” is when a startup finds a large and sustainable group of customers who resonate with and express loyalty for the product (219). The chapter combines the “product/market fit” concept with the engine of growth. The engine of growth can be used to measure how far or close a company is to the “product/market fit” by evaluating the data from the Build-Measure-Learn feedback loop. Lastly, every engine of growth runs out eventually. The successful startup will recognize early when the engines of growth are slowing down and make the appropriate pivots to continue sustainable growth. Past successes cannot alone dictate the future of a company’s product development, marketing, and business strategy.

To avoid the stagnation described at the end of Chapter 10, companies need to become adaptive organizations. Chapter 11 describes how companies can manage portfolios while tuning the engines of growth and finding new areas of development when customer interest or activity starts to slow. The first step in building an adaptive organization is developing a training program for new employees. The process of orientation needs to constantly align with the vision of the company and evolve with the external factors affecting growth.

Before diving into the details of adaptive organizations, the chapter addresses a perceived paradox in his own message. On the one hand, The Lean Startup advocates for validated learning and MVPs that produce product for the customer’s purchase and use at a rapid rate. On the other hand, the Build-Measure-Learn feedback loop stresses a prolonged continuity and analysis of the product development process. Inevitably, the relation between innovation and feedback causes interruptions and slows a startup down. The adaptive process of an organization embraces these slowdowns as part of the natural feedback loop. For instance, implementing a new training program for employees will slow down the work of other employees, but the improvement in speed after the training, if executed effectively, will increase productivity. The system for new employee training is called the “Five Whys.”

The “Five Whys” is a method addressing and preventing the most problematic issues in a company. By asking “why” five times, the team can investigate the root of problems in systematic fashion. One example is from Taiichi Ohno, a founder of the Toyota Production system:

• Why did the machine stop? (There was an overload and the fuse blew.)
• Why was there an overload? (The bearing was not sufficiently lubricated.)
• Why was it not lubricated sufficiently? (The lubrication pump was not pumping sufficiently.)
• Why was it not pumping sufficiently? (The shaft of the pump was worn and rattling.)
• Why was the shaft worn out? (There was no strainer attached and metal scarp got in.) (230).

The repetition of five “why” questions seeks to identify the root problem so that the team can fix it. Importantly, the “Five Whys” method begins with a technical issue but almost always ends in some human error. Using this method, Ries employed his own “Five Whys” to start a company training program:

• A new release disabled a feature for customers. Why? Because a particular server failed.
• Why did the server fail? Because an obscure subsystem was used in the wrong way.
• Why was it used in the wrong way? The engineer who used it didn’t know how to use it properly.
• Why didn’t he know? Because he was never trained.
• Why wasn’t he trained? Because his manager doesn’t believe in training new engineers because he and his team are “too busy” (231-32).

The Five Whys method can be used to build an adaptive organization by investing time and attention at each level of the company hierarchy. This also generates accountability from the top down.

The Five Whys approach also regulates the speed of a company’s problems and solutions. By connecting the questions to learning outcomes, the Five Whys can organize a team in a way that manages and overcomes failures more efficiently. Yet, when the Five Whys goes wrong, what Ries calls the Five Blames, company morale and productivity plummet because team members start blaming each other. To escape the Five Blames, companies need to ensure that everyone affected by or who encountered the problem in question is present in the discussion about it, from customer service reps and engineers to managers and executives. During the root cause analysis, blame inevitably starts to fly, so the chief executive should remind the team that “if a mistake happens, shame on us for making it so easy to make that mistake” (235). The Five Whys is a systems-level protocol for identifying problems and solving them, not an opportunity to assign blame or make excuses. Companies who adopt the Five Whys should be tolerant of all mistakes in the initiation period and never allow mistakes to be made twice (236).

The Five Whys can cost the company time and money. Because of the deep questioning method, unpleasant facts about a company’s organization and culture will arise. The key to managing this tough initiation period is to start small and be specific. Instead of applying this system’s analysis approach to everything in a company, the team should address small issues first to understand the process of addressing the problem with the Five Whys. After everyone has voluntarily agreed to practice this method, the team should appoint a Five Whys master who can act as a moderator for the meetings. For example, the Five Whys could be applied to problem with a new feature rollout from a small batch of product. Testing out the methods on a small and specific issue can alleviate the growing pains of using the system’s analysis approach.

Ultimately, an adaptive organization needs to efficiently move through the Build-Measure-Learn feedback loop. A lean technique, like the Five Whys, helps retain employee commitment and assists in training new employees on the lean manufacturing principles.

Chapter 12 discusses how startups can continue to address the needs of existing customers while exploring new innovations and markets with a term called, “portfolio thinking.” Startups possess resources far less than established organizations. Their capital and time to succeed are finite. Conversely, because of their lack of resources, most startups have independent authority over their business strategies and a personal stake in the outcomes. The problem with this reality is that the parent organizations or investors of a startup hold fear about the startup’s potential successes and failures. These structural attributes of the startup can be manipulated to change the management philosophy and nurture disruptive innovation.

Startups need to operate with a management philosophy that protects the parent organization, holds entrepreneurial managers accountable, and reintegrates innovation back into the parent organization, if successful (256). The chapter recounts an issue at a company when different departments failed to agree on the interpretation of crucial data. Every department left the initial meeting with separate conclusions about the problems, and potential solutions, facing the company. For context, the organization served a business-to-business segment and a consumer segment. The business-to-business drove much of the revenue, but growth had slowed. The parent organization agreed that the consumer segment held potential for growth, but nothing was materializing. Out of a fear that experimentation would disrupt the revenue on the business-to-business segment, the company failed to develop solutions to grow the consumer segment. In sum, the parent organization within the company feared that the intrapreneurial team on the consumer segment side would fail and subsequently tank the entire company.

This example introduces the concept of an innovation sandbox, a process of experimentation that cultivates innovation without sacrificing the growth or revenue of the parent organization. In brief:

• Teams are allowed to create split-test experiments that only affect the sandboxed portions of the product or service.
• The team must complete the whole experiment from beginning to end.
• Experiments cannot run longer than a specified amount of time.
• Experiments cannot affect more than a specific number of customers.
• Each experiment is evaluated on a single standard report of actionable metrics: “five to ten (no more).”
• Each team in the sandbox uses the same metrics to evaluate the product’s success.
• All teams that create an experiment must monitor the metrics and customer feedback in progress, and immediately abandon the experiment if catastrophe occurs (261-62).

This is another manifestation of small batch manufacturing, albeit in a controlled, experimental form. The sandbox method allows teams to innovate and learn from failure while alleviating the fears of disaster from the parent organization. The results of the sandbox should be collected into an overall portfolio of company activity and growth. All internal teams should have access to this portfolio so they can understand the products and services in development and on the market.

One problem that occurs in startups who work with this method is that the innovators end up monitoring the product after it reaches the commercialization phase. This is a waste of the employee’s and company’s valuable and scarce resources. Instead, products should be handed off between teams to highlight the strength of individual employees. “People should be allowed to find the kinds of jobs that suit them best” (266). Here, companies should create a new job title for innovators within the organization: “Entrepreneur.” To avoid innovators/entrepreneurs/intrapreneurs getting stuck in the status quo, the product should be handed off to a larger team after the initial experiment becomes a success. By forcing teams to work cross-functionally, the startup can achieve validated learning throughout the steps of product development and marketing. Lastly, it is helpful to work through the growing pains of learning the Lean Startup method by practicing in a community of like-minded individuals. 

The Epilogue to The Lean Startup recounts the 100th anniversary of Frederick Winslow Taylor’s, The Principles of Scientific Management (1911), which created the discipline of management science implemented in American corporations. The notion that organizations and their operations could be studied as a science may be credited to Taylor’s work on the subject. Unfortunately, some of Taylor’s supporters interpreted his theories incorrectly and started to advocate that “workers should be treated as little more than automatons” (273). Lean manufacturing seeks to remedy the harms of Taylor’s theories by reimagining his “core idea that work can be studied scientifically and can be improved through a rigorous experimental approach” (273). In the 21st century, organizations face new challenges, unimaginable to early 20th-century thinkers. Modern productive capacity exceeds the awareness of what to build because nearly anything can be produced. Despite modern society’s ability to feed, clothe, and shelter nearly all of humanity, the productive capacity of the world continues to waste efficiency and resources, ultimately leaving these aspirations unfulfilled.

The Lean Startup method seeks to eliminate most of the waste in innovation by attempting to learn the unknown. The scientific method can advise entrepreneurs on how to build sustainable organizations around new sets of products and services (275). However, individuals are still the most important part of the process. The adaptability, creativity, and wisdom of individuals can enhance the efficiency of empirical research and business systems. The traditions of entrepreneurship—vision, risk-taking, and courage—can be supplemented by the systematic approach to building innovative products.

The chapter warns against turning the Lean Startup movement into pseudoscience: “Any time a team attempts to demonstrate cause and effect by placing highlights on a graph of gross metrics, it is engaging in pseudoscience” (279). The same criticism applies when a team justifies failures as learning without empirically tracking the feedback from experiments and customers. Moreover, the audience should not throw around the language of pivots, validated learning, and MVPs haphazardly without understanding their core concepts.

The end of the Epilogue advocates for a new research program that could discover how organizations could work more efficiently. For instance, researchers could learn what “stimulates productivity under conditions of extreme uncertainty” (281). Startup testing labs should be created that experiment with product development methodologies. These labs would bring in small, cross-functional teams and have them solve problems using different development methodologies. Competitions could be opened to raise the stakes of each team’s development efforts, like the international programming competitions that have developed databases with solutions out of complex problems (281). Each competition could increase the level of uncertainty in a given product development situation. Real-world customer problems would be used as prompts to engage the teams’ creative and innovative thinking. Additionally, teams could try to develop MVPs that could solve a certain problem. All of these possibilities relate to Ries’s prediction that new research will make the lean startup method a center of entrepreneurial practice in the future.

The last section of the book describes Ries’s newest project, the Long-Term Stock Exchange (LTSE). The LTSE is a new type of exchange that trades the stock of companies devoted to sustaining long-term thinking. Companies would report using innovation accounting to track their internal entrepreneurial efforts: “Executive compensation in LTSE companies would be tied to the company’s long-term performance” (283). LTSE companies would be given the freedom to manage and pursue long-term investments and support long-term thinking for product development strategies. Lastly, LTSE data would be made transparent so that others could learn from the data and start to innovate at their own organizations.

To Ries, The Lean Startup “avoids doctrines and rigid ideology” (283) and can unlock human potential. All creative hypotheses would be tested with rigor. Speed and quality would be allied in the pursuit of the customers’ benefit. Vision would be consistently tested and verified. The response to failure would be faced with honesty and empirical research so that speed and learning grow. Sustainable growth and value would supplant wasteful aspirations and efforts. Most of all, the Lean Startup “would stop wasting people’s time” (284).