CHAPTER 7
SELF-TUNING
How to Make Strategic Processes Smart
One of the most important organizational processes a smart business needs to change is how it formulates and implements strategy.1 Strategy no longer means long-range planning. It’s not even short-range planning. It’s not planning.
Fundamentally, strategy making is now a dynamic and fluid process, akin to learning. The classical approach of analyze, plan, and execute is much too slow and inflexible for today’s environment. Instead of formal planning, strategy formulation is the constant and rapid iteration between vision and action. Constant experimentation creates feedback, which leads to adjustment of the vision, which in turn guides new experiments. Strategy is constantly updated in this iterative process. With the right infrastructure and leadership, this strategic experimenting process becomes core to the nimble, innovative, smart business.
In traditional strategic planning, management makes trade-offs between exploration and exploitation: both have costs that must be managed. In the most complex, dynamic environments, exploration and exploitation must be done in parallel and continuously, without interruption for assessment. Through strategic experimentation, a business must constantly absorb new information and test ideas and processes to adjust its strategies to the new reality and opportunities. Fortunately, with new technologies and infrastructure, the cost of experimentation has been dramatically lowered. But you must have your organization and strategy in place to experiment on a large scale. Your enterprise must be able to constantly adjust to new ideas and changes in the environment rather than simply manage them in the traditional fashion.
Learning and innovation have long been ideals of business development. Indeed, the very phrase learning organization has been around for roughly three decades. But, like customer-centricity, the learning organization has been more of an exception than a rule. Incentives and limitations inherent in the industrial-era model work against learning. Enacting a learning-organization culture in a traditional hierarchical organizational structure focused on execution and minimizing transaction costs is extremely difficult. The US military, for example, has long understood this limitation and inability to absorb and act on local knowledge and has struggled with various work-arounds to remedy these problems.2
Despite new technologies that have lowered coordination costs dramatically, made information transfer immediate, and automated some testing regimes, strategic planning has remained largely the same. Strategic planning departments provide strategic options and execution plans that are based on static analysis, and then management chooses the strategy. This process is ill-adapted to today’s environment. Besides being slow and inflexible, traditional strategic planning makes little use of the data and machine-learning resources both inside and outside the business. These capabilities can speed and amplify the effects of strategy. If an enterprise datafies all aspects of customer interactions and partner activities, it can see the results of its experimentation, such as A/B tests, in real time. Then machine-learning algorithms can automatically make adjustments that increase systemwide efficiency and make exploitation of a successful strategic experiment efficient and even semi-automated.
Strategic planning departments or their successors can now focus their efforts on developing creative prototypes of products or processes—the exploration process—that can be fed into further experiments. I call this experimental strategic cycle “self-tuning,” in an explicit nod to algorithmic design thinking from chapters 3.Applying self-tuning thinking to strategy and even to the organization is no easy task. This chapter shows how Alibaba has been making this transition.
Dynamic Strategy: Strategic Adjustment in Real Time
In 2008, Alibaba identified its strategy for the next ten years as “fostering the development of an open, collaborative, and flourishing e-commerce ecosystem.” But it wasn’t until the last few years that we started realizing that in responding to the changes in the external environment, the organization itself had to evolve. We didn’t know enough about what the future would look like a few years out to plan for it. Alibaba needed to constantly adjust and readjust to the environment in real time, without traditional management getting in the way.
Self-tuning means that learning becomes the central focus of an organization. The strategy-making process is one of generating, coordinating, and modifying experiments—a vastly different operation from traditional long-range planning. The organization seeks a coherent vision of the future, both in goals and in execution. It implements this strategy by experimenting on that vision throughout the business. When vision and the experiment intersect, success is imminent. At Alibaba, we have made great efforts in this direction, including studying what other internet pioneers have done. Our multiple businesses and continuing growth testify to our success in these early stages of experimentation, though there is much more to be done.
The Self-Tuning Learning Loop
Machine-learning algorithms are a productive analogy for a self-tuning organization; they embody learning loops that prompt self-adjustment. I discussed in chapter 3 how datafication, iterating algorithms, and smart products combine into the capabilities of data intelligence. Thanks to data intelligence, MYbank’s loan product learns from borrower behavior, and the business continuously improves its lending decisions. The business’s algorithms continuously update their lending rates according to feedback in the form of data such as loans accepted and time to repayment. MYbank’s algorithms improve recommendations over time—their goal is to decrease the total default rate of the platform as a whole. (In computer science, the optimization goal of an algorithm is called the “objective function.”) In chapter 3, I analyzed MYbank’s product with a focus on how the lending business developed data intelligence. Now, I will extend the basic mindset behind data intelligence into a larger feedback loop, which organizations as a whole can apply to strategy formulation.
Adjusting through experimentation
Machine-learning algorithms are designed to generate, test, and amplify favorable outcomes. They exhaustively sift through an array of possible options; in MYbank’s case, these options encompass all the possible variables that might affect a lendee’s repayment. All of these possible outcomes represent hypotheses. Is default rate connected to the time of day when the loan is issued? Will this batch of users default? Machine-learning algorithms conduct experiments on enormous data sets, testing new ideas and recording the result. As the business grows, datafication continues and generates more information, with which the algorithms can generate even more options and experiments.
Experimentation does not occur at random. Engineers design algorithms to test various options economically, minimizing search time and computing cost. One of the chief techniques for economizing experimentation is to amplify what has worked in the past and replace less preferred options with different choices. Experimentation levels are moderated after machine-learning engines learn more and more about users through sustained interaction. MYbank’s algorithms hone in on the risk tolerance of the lendee with each iteration, lowering the rate of recommendation of randomly generated products. The algorithms then apply this information to further experiments to amplify what worked.
As experimentation starts to converge on a result, the customer experience begins to improve as the adaptable product adjusts. MYbank’s lending services get better and more responsive as they scale. It is no exaggeration to say that the algorithm engines shape, to some degree, user experience and even user behavior, in the case of more complicated data-intelligence products such as recommendation engines on the Taobao app. Much of the user delight with recommendation engines comes from finding new products and content they would not otherwise have found. Being directed to a new category or product both uncovers and shapes what a user finds interesting and what he or she will look for in future visits to the marketplace.
Taken as a whole, the experimentation process of machine-learning algorithms can be called “self-tuning.” Self-adjusting functionality is baked into the algorithms themselves. No analyst or programmer needs to individually interpret user feedback, manually adjust the exploration ratio, commission an analysis, or deliberate on how best to guide users to more optimal behavior. Data intelligence takes care of much of the hard work, helping businesses adjust independently to a wide variety of environments, especially ones that change constantly. The most developed data-intelligence services can conduct experiments and make adjustments for each user in a rapid and massively parallel fashion.
Applying Self-Tuning Principles to Strategy
Products built on data intelligence are self-tuning, navigating with minimal human input though the complex, uncertain, and fast-changing landscape of user desires and requirements. But these products also herald a new trend that is often overlooked even by those quite familiar with artificial intelligence. The world’s top internet companies, such as Facebook, Amazon, and, of course, Alibaba, are taking the self-tuning mindset behind data intelligence a step further, experimenting with how self-tuning principles can be applied to the entire enterprise.
To imagine how we might run a firm like an algorithm, recall that computer algorithms do not program themselves. Humans must decide their objective function, the overarching goal of the algorithm, and adjust how the algorithm prioritizes different directions toward that goal. (For example, the objective function of Facebook’s news feed product is a combination of advertising revenue and user engagement, measured by metrics like the number of comments posted. By changing the relative weight of these two metrics, Facebook can make effective trade-offs between user experience versus income. I will return to Facebook’s goal metrics in the next chapter.)
In an organization, the analogue to objective function is an organization’s vision: as vision changes over time, the business model will evolve. As the rest of this chapter explains, the self-tuning characteristics of algorithms can shed light on how to structure and run whole enterprises in an increasingly complex and fast-changing business environment. As the organization collectively monitors the competitive environment, consumer engagement, and system-wide results, it intervenes when necessary, adjusting infrastructure, goals, and vision to shape a healthy, productive business. Such an organization is built for change and emphasizes planning and experimentation, not finalizing plans.
The Starting Point: A Vision of the Future
In most organizations, the vision and business model are fixed axes around which the entire enterprise revolves. They are worked out by the founder or founders, and once proven successful, they are rarely changed. The organization is essentially focused on realizing this fixed vision by executing, optimizing, and scaling the foundational business model.
In the traditional industrial economy, the vision or mission and the business model were often the same: build cars affordable to the masses, connect the country by railroad, provide electricity, and so forth. The vision or mission was less important when the future was more predictable. An organization’s vision has progressively become more important as business makes the transition to a knowledge economy and as economic and technological change has sped up. Senior management needed a way of guiding employee actions, communicating goals with investors, and applying institutional knowledge to more markets and products. So, they started crafting a vision. They then generated new ideas to improve the current offering or to create new offerings, while still firmly situated within the frame of the existing business model and vision.
This traditional approach is clearly self-limiting, not only in a world of fast-changing customer preferences and offerings, but also in one of shortened enterprise life and rapidly aging business models and strategies. Today, vision is the heart of a firm’s strategy. It must be clearly understood, trenchantly described, and regularly updated to bring together a network of suppliers, producers, partners, and customers. Vision sets the direction for the evolution of the whole network. Like customer first, vision has changed from nice-to-have to an operational must-have. It is the objective function for the organization and network.
I will discuss mission and vision in much detail in the next chapter, but for now, it is important to explain the difference between the two terms as Alibaba understands them. At the highest level, vision is an understanding of the world as it will be in the future. It captures the direction industries will evolve in response to societal, economic, and technological progress. Only on the basis of that understanding, within a landscape of change, can the firm articulate its direction and ambitions. Thus, vision, narrowly written, describes what the firm seeks to achieve and defines the scope of exploration. It defines where the firm fits into the future.
The mission of the firm, on the other hand, is the change the firm is driven to make in the world. It is the firm’s reason for existing, and the clarion call by which it attracts talent and resources to its side. Mission and vision (as well as values, which I will discuss in the next chapter) are closely linked and influence the other: your view of how the world will change will necessarily affect how you might change that world. For some companies, mission and vision are treated as the same thing. However, I would advocate separating the two to achieve a balance between a relatively fixed raison d’être (mission) and a mutable, improvable view of the future (vision). Practically speaking, a successful firm does not normally iterate on its mission the way that I advocate iterating between vision and action in this chapter. (Alibaba’s mission, as I will describe in the next chapter, has basically remained constant throughout the life of the company: “to make it easy to do business everywhere.”)
This chapter will focus on vision and how to steadily improve it through experimentation. Today, as I’ve described in this book, organizations are but one player in an interconnected, intelligent, evolving network. Whoever has the vision for the future will attract players into their network—the partners, providers, and consumers. The more connections the vision inspires, the more assets the visionary can mobilize. It is the only way to shape the future around you.
Why visionaries dominate
Jack Ma, Steve Jobs, Elon Musk, and Mark Zuckerberg are all known as visionaries. They have to inspire their employees, partners, and customers to mobilize the network to realize the vision. They are outspoken evangelists in a way that the leaders of GE, Toyota, and Merck have never been. The difference in the demeanors of these two groups is not coincidental.
Traditional corporate leaders do not shape the future; they run a machine. When we teach at Alibaba’s School of Entrepreneurship, we show a slide of ten business leaders and ask our students to identify them.3 They can easily pick out Ma, Musk, and Jobs. Communicating a vision to many people grants a form of celebrity and often requires a highly individualized value set and mission. But virtually no one can identify the CEO of Citibank, Toyota, or General Electric. Those companies don’t need visionaries. They need executives who can manage operations. Such leaders are much more interchangeable.
In 2015, Lyft founder and CEO John Zimmer described to me how, when he started the company’s ridesharing service, he worried about its ability to gain traction. Unlike other rideshare services, this one tried to connect several riders to minimize car use and to be more eco-friendly. But without a lot of riders, early adopters would have much longer trips and save little money. Zimmer was shocked at how quickly the service ramped up. Plenty of customers shared his vision and were willing to make short-term sacrifices to make it happen. In classic non-zero-sum fashion, as the number of customers increased, the service became more efficient and useful.
The web celebs described in part 2 are more than successful brand builders. These entrepreneurs continually demonstrate new visions of themselves, their looks, or the environment. More importantly, despite being just one individual, web celebs like Big-E have a vast network supporting them. Firms like Ruhan are visionaries within the wider network, shaping the future for themselves and for their partners as the apparel industry realigns around them.
Retuning the Vision
A clear vision shapes the future and directs the network’s evolution, but the vision must continually incorporate feedback and evolve within the larger environment. More than a static vision, the firm needs a visioning process. As time passes, its vision has to be checked against reality and updated.
In a network, the boundary between the internal and external organization is blurred. More importantly, as I described in chapter 6, the strategies of different firms are interdependent, especially for platforms. Information from the environment must be taken in to retune the platform’s vision. In this dynamic process, leadership needs to continually reshape its understanding of the future, thus influencing the future shape of the system. Steve Jobs used Apple’s product launches, always spectacular affairs and self-conscious theater, to express his revised vision. That’s why it is called Macworld: the event is an essential but often-overlooked way in which Jobs “managed” the Apple ecosystem.
Alibaba exemplifies this revisioning approach. When the company started in 1999, the internet reached less than 1 percent of China’s more than one billion citizens. While many observers expected penetration to grow, they couldn’t predict the precise nature of that growth. In response to this uncertainty, Alibaba applied an experimental approach to our vision. Rather than treat our vision as a given, the company instead posited a vision, given the best working assumption about the future and using all available information. We remained transparent about this approach. As the market evolved and new realities emerged, management regularly and profoundly reevaluated its vision, checking its intuition against reality and modifying the company’s goals as appropriate.
In the early years, Alibaba directed its efforts to becoming, in the company’s words, “an e-commerce company serving China’s small exporting companies.” This objective led to the initial focus on Alibaba.com, which created a platform for Chinese manufacturers to sell internationally. However, as the market continued to evolve, so did the company’s vision. With the explosive growth of Chinese domestic consumption, Jack Ma saw the opportunity to expand our e-commerce offerings beyond China’s export businesses to include Chinese consumers. The result was the launch of Taobao in 2003. However, Alibaba soon realized that Chinese consumers needed more than just a marketplace for buying and selling. They needed greater confidence in online shopping and assurance that their payments were safe. There were no credit cards in China at the time. Consequently, Taobao expanded its reach with Alipay in 2004, which became a runaway success and greatly sped up the penetration of e-commerce across the country.
We at Alibaba were not the only ones to create services to facilitate the marketplace. As described in chapter 2, other companies providing services, such as storefront builders and models, entered the growing marketplace. Building on this development, Alibaba expanded its vision in 2008: “Foster the development of an open, collaborative, and flourishing e-commerce ecosystem.” The company started to offer more infrastructure services, such as cloud computing, finance, and logistics. With the rapid emergence of the mobile internet, Alibaba has further evolved our vision to its current version: “We aim to build the future infrastructure of commerce. We envision that our customers will meet, work and live at Alibaba, and that we will be a company that lasts at least 102 years.” Note that we deliberately dropped the e from e-commerce, to reflect Jack Ma’s belief that all business will become e-business. (For much more detail on the strategic choices made throughout Taobao’s history, see appendixes A and B.)
Many outsiders view our vision statements as PR exercises that help the world understand the firm, not as actual descriptions of our business model. Nothing could be further from the truth. Alibaba’s evolving vision reflects our top management’s understanding of the future of commerce, as well as Alibaba’s place in constructing that future. Only with the current vision articulated can the entire network and our organization begin to move toward that future.
Dynamic Strategy: Planning, Not Plans
As vision the static noun is inflected into the active verb of visioning, the strategic process must emphasize the dynamic. It is about the planning, not making a plan.
In many organizations, detailed, fixed plans form the core of their strategy. In Alibaba’s case, rapid technology changes, shifting consumer expectations in China and beyond, and regulatory uncertainty make it very difficult to predict or plan for the future. In response, we have shifted the conventional focus on plans to a continuous process of planning. Rather than an elaborate plan that is executed meticulously, Alibaba continuously retunes strategy as circumstances change, in a very decentralized manner.
Within the company, there is a normal annual planning cycle, with a few iterations between business-unit leaders and the top management team in the third and fourth quarter each year. Leadership, however, recognizes and expects that this direction is only a starting point and will change. In my nearly decade-long capacity guiding the strategy team, I have never written a formal strategy plan for the company. But each year, we do have a ten-page presentation summarizing the key points of our strategic understanding. This presentation is often reviewed, and revised if necessary, whenever we discuss important business issues.
I remember an animated discussion at a top management meeting one day in April 2012. We reached consensus about the importance of data in the future and agreed that Alibaba should become the leading platform for data sharing. Later that same day, there was a scheduled meeting that included Alibaba middle managers. Jack Ma asked me to summarize the morning discussion during lunch and present the new ideas as our new “plan” to the whole group that very afternoon. This mindset might strike an organization accustomed to spending several months to create a three- or five-year plan as slightly unorthodox. For us, it is par for the course.
Moreover, such rapid iterations of strategy happen at every level of the organization. Whenever a business leader sees an important change or a new opportunity in the marketplace, he or she can initiate what we call internally a cocreation (gong chuang) meeting. It is called cocreation because Alibaba employees, including senior business leaders and lead implementers, codevelop new directions with customers.
Cocreation
Cocreation, typically kicked off with a full-day working session at Alibaba, involves four steps. First we set the “ground truth,” identifying and laying out change signals based on data from the market or insights from customers or staff. In addition, we ensure that the right people are in the meeting and have the right dynamics to work together toward a solution.
Second, we get to know the user and their current situation in as much depth as possible. In this step, the participants dive deep into the merchant or consumer viewpoint to understand our users’ evolving needs or pain points and to brainstorm potential solutions. For example, in a recent cocreation meeting for one business unit, Alibaba selected five consumers to join the meeting and divided the staff into smaller teams to each work with one consumer to understand their pain points. The teams then reported back to the larger group twice—once with the user present to ensure that the issues were understood correctly and the second time without the user to push to a deeper level of analysis and solution development.
Third, we base an action plan on the outcomes of the discussions. The action plans must identify a leader who can champion the issue or opportunity, the supporting team or teams that will put the ideas into action, and the mechanism for doing the work. The mechanism involves, at a minimum, the communication processes, the metrics for evaluating progress, and the timeline for execution so that the team can align its efforts. Often, this third step takes the most time and effort, because action plans are truly crucial for ensuring that ideas gets translated into reality. Figuring out who is in charge of making the action plan happen is a complicated and often-fraught process, but teams need to do the hard work to get results.
The fourth and final step of cocreation is user feedback. Teams must embed regular feedback into their development processes. At monthly or bimonthly check-ins at Alibaba, teams present user reactions to evolving designs, prototypes, or concepts to ensure that execution meets market expectations.
Taken together, the steps of cocreation highlight the iterative and distributed dimension of self-tuning. Business units can initiate cocreation sessions when they see a relevant market stimulus, without any central mandate or oversight. By creating a forum for regular exchange with customers and related parties within the company, Alibaba evolves with the market and makes optimal use of local knowledge. In effect, senior leadership relinquishes a degree of control to allow the enterprise to organically adapt to the external environment. In this way, self-tuning enterprises leave more room for products and even business models to be pulled by the market, rather than pushed via top-down decisions. All this effort also contributes to the continuous upgrading of the vision or strategy at the senior level, as decision making and its repercussions “trickle up” through the organization.
In summary, there are three key points in the strategy-making process for smart businesses. First, vision is critical. Second, smart businesses constantly retune their visions. Third, for smart businesses, strategy formulation is dynamic; it is about planning, not about static plans. What replaces traditional strategy formulation in a smart business is iteration through experimentation.
Applying Self-Tuning Principles to Business Models
When the vision and strategic plans are no longer fixed points but dynamic processes, so is the business model. Indeed, the business model as a whole is the most important arena for experimentation.
Experimenting with the Business Model
Previous chapters have discussed many examples of why Alibaba’s ecosystem consists of much more than retail in the traditional sense, from e-commerce platforms (Taobao, Tmall) to payment and finance (Ant Financial), cloud computing (Alibaba Cloud), and logistics (Cainiao Network). (See appendix A for more details on these businesses.)
To achieve such breadth and depth, we had to commit to business model experimentation from the outset. The decision in 2003 to venture beyond our core B2B business and found Taobao—just four years after our company’s founding—was hotly debated within the company. A formidable rival, eBay, had already entered the Chinese market with much fanfare, but to the leadership team, the American-based company seemed to be operating in several ways out of step with the Chinese market. However, Alibaba did not give up on exploitation of our budding B2B business. To minimize downside drain, management set up the new venture as a startup, with separate funding. (In fact, Taobao was a fifty-fifty joint venture between Alibaba and Softbank.) The Taobao team spent its early days in an apartment, totally separate from the Alibaba offices, and experimented as it saw fit. By sequestering the new venture and its employees, Jack doubled down on experimentation.
At each juncture of its growth and evolution, Alibaba generated new business-model options, testing possibilities by letting them run as separate units. The most promising ones then scaled up. In 2006, seeing two new developing trends, B2C and SaaS (software-as-a- service), we started two new business units to experiment in those spaces. The Taobao Mall, which after a few iterations of the business model became Tmall, is a major part of the group portfolio today. On the other hand, AliSoft, which tried to catch the SaaS wave, entered the market too early and couldn’t find a killer app with enough customers. The business was shut down in 2009.
Focusing on Exploration
Another driver of Alibaba’s success has been its deliberate choice to keep experimenting with the business model in response to the environment. Rather than being content to transition into exploitation once the business model matures, Alibaba continues to engage in exploration as new conditions emerge.
For example, Taobao achieved more than an 80 percent share of China’s e-commerce market within just four years of launch and became a national phenomenon by 2011. Many would take this leadership position as a signal of market validation and focus on optimizing and defending the successful model. Instead, we saw the still-unrelenting growth of China’s internet population and the increasing sophistication of consumers and retailers as a signal of greater uncertainty in the marketplace and a risk to our current model. This is where human judgment comes in. Deciding when to keep exploring or switch to pure exploitation will define the culture and success of many businesses.
For Alibaba, leadership was inclined to conduct more experimentation and take more risks. Again, there was heated debate within the company on which direction to go and which business model to build up. Instead of making a simple top-down decision, we made a bold experiment: let the market pick the future winners. In 2011, Alibaba split the successful Taobao business into three independent and competing businesses units. Each unit would effectively make a different bet on the future of e-commerce in China. Taobao would focus on smaller brands and the consumer-to-consumer (C2C) market; Tmall on larger brands and the B2C market; and Etao, a new business unit, would focus on product search, aggregating information across different marketplaces and platforms.
Increasing your experimentation at the height of success runs contrary to established managerial wisdom, but for Alibaba, it was a necessary move to avoid rigidity and to continuously create options in a rapidly evolving e-commerce market. By early 2013, Tmall had won market leadership in fiercely competitive B2C markets—a successful experiment. Taobao maintains its dominant position in the C2C marketplace and has since given rise to innovative C2B businesses such as the web celebs—another successful experiment. Product search, on the other hand, proved not to be the future, and Etao has become a niche product.
Obviously, such experimentation comes with high financial and organizational costs. I remembered the tremendous pressure we faced when we split Taobao into three units. It was very hard to tell employees that while they were competing against each other in the market, they also belonged to the same company. But being straightforward about what we were doing was also very important. Employees needed to know that the experiment was taking place and what we were trying to learn. Discerning the real drivers of each business’s development is also very difficult. The time spent setting up, communicating, and letting the businesses duke it out might have been inefficient, but the costs are worth it. In a rapidly changing environment, getting your vision of the future right and securing the fit of your strategy with the evolving environment are the most important objectives. The significant investment in experimentation is well worth the cost and the risk. For Alibaba, our experiments clarified the direction of e-commerce in China and kept providing the resources that fueled the growth of Taobao and Tmall—as exemplified by our stunning Singles Day successes.
Returning to the Vision
Through business-model experimentation, Alibaba has not only arrived at a clearer vision of the dynamic environment in which we operate. Our vision has profoundly shaped the evolution of our environment.
For smart businesses, leadership and the entire organization should be structured to experiment and report results, even those that are unsuccessful or have unintended consequences. The organization must be minutely attuned to its network and market. The vision, the objective function of the “algorithm” that is the firm, may need to be recalibrated, improved, or changed altogether. A vision, just like algorithms, needs to be guided and modified by constant human inquiry to make sure both customers and the overall ecosystem are evolving in healthy ways. Recall from chapter 4 how Alibaba kept adjusting its search algorithms to promote a balanced and robust market for both buyers and sellers, and how our recommendations product changed after the 2013 transition to mobile.
As our vision of the future changed, our platforms set the direction of China’s e-commerce sector. Tmall has evolved from a higher-end marketplace on Taobao to the entrance point for the world’s global brands to enter China through Tmall Global. Taobao has come a long way from its origins as a digital flea market; it now enables consumers to shop for anything imaginable, and incubates highly innovative firms such as the web celebs. At the same time, Alibaba’s enabling business models, such as Ant Financial for finance and the Cainiao Network for logistics, have set new expectations for security and computing online and offline. Our dominance today is a result of consistently improving upon our vision, and then daring to allow our businesses to capitalize on a new and improved view of the future.
The Foundation as a Change-Seeking Culture
Faced with market fluctuations and disruptions that regularly reshape the landscape, an organization with a self-tuning strategy sets aside the idea of a fixed vision and business model. Instead, the organization regularly recalibrates all its components to the environment by continuously experimenting throughout the organization. The goal is that the firm’s vision will begin to converge through a process of action and recalibration—in other words, self-tuning. Self-tuning thus implies that the organization is always learning and innovating. Consequently, change is the natural outcome and an essential feature of the organization.
Setting Expectations from the Beginning
With all this retuning and experimentation, a culture that not only facilitates but even encourages change is fundamental. Reactions to change depend heavily on the mindset of the organization. At Alibaba, the embrace of change is wired into its DNA. The company has created a language and an expectation of change through its six core values, one of which is precisely “embracing change.” Jack Ma regularly emphasizes this theme in his communications with employees, as does the rest of the leadership team. Leadership is totally forthcoming with employees about anticipating change and adaptation at every level from the day that new employees join the company.
By creating the expectation of change, Alibaba’s employees have come to see it as part of business as usual. “If you have not changed bosses five times in a year, you haven’t seen real change” is a well-circulated adage. Customer first is the first of Alibaba’s six core values. Since customer needs are a moving target, you have to change to meet their needs. Otherwise, you will simply be wiped out by competition. “To be prepared for change,” Jack Ma emphasizes, “is the best plan.” You have to evolve with the external environment, no matter how fast you have to do it, and change must be hardwired into your organization’s culture.
To build this culture, you need the right people. A key consideration in Alibaba’s hiring decisions is a candidate’s willingness to change. Experience has shown that technical skill alone is not a sufficient metric to identify the right talent. Rather, when assessing candidates, interviewers regularly ask about the biggest change a candidate has experienced and how he or she dealt with it. In this way, Alibaba only brings on board new employees who are ready and able to change. I will expand this point in detail in the following chapter.
Institutionalizing Change
Organizational change must be institutionalized and normalized. To be good at change, an organization must regularly engage in it. Traditionally, organizational change is conducted through infrequent, but major transformations. However, if an enterprise regularly adjusts itself to the external environment, it has less need for risky one-shot overhauls.
An extreme example occurred in 2012, when Alibaba experimented with a rotation program for its twenty-two main upper-level managers across the business. While rotational programs are not uncommon at more junior levels, Alibaba focused its program on the senior-most levels of each business unit (i.e., shuffling around all leadership except the C-suite). There was some concern about the potential risk this program would bring to continuing operations. However, the program proved quite successful, as it required managers to institutionalize and transfer knowledge to transitioning colleagues, thus preventing siloing and parochial thinking. The program not only helped further develop the skills of top talent, but also demonstrated throughout the entire organization the leadership’s commitment to organizational flexibility. Alibaba now runs a regular program rotating a portion of senior leadership every year.
Alibaba continuously focuses on developing and maintaining an organizational flexibility that tracks the environment, especially as the company continues to grow. As Ma has explained many times in internal meetings: “Strategy and organization go hand-in-hand. Every year we change the organizational structure in tandem with changes in strategy.”
I mentioned above how Taobao was split into three business units in 2011. Two months later, the whole group was first split into seven business units and then further divided into twenty-five business units in the following sixty days. The objective was to make the whole company as nimble as possible, so that each business unit could move quickly on its own. Remarkably, after three major reorgs, each business unit finished its strategy, annual planning, and budgeting process in three months. Then, in late 2013, Alibaba launched an all-in-mobile initiative to become a mobile-first company. In addition to normal reorg, the company drafted 5 percent of engineers from all business units—often some of the best—and moved them to the mobile initiative. Though there were growing pains, most employees understood that these changes represented a conscious effort on the part of the company to adapt to the environment and to prepare for the future. All this would be impossible without a culture that embraced change.
Many people talk about experimentation, innovation, and organization change, but the cost is simply prohibitive. Internet companies like Alibaba can do this because they have built the right culture and the right infrastructure. I will elaborate on culture and infrastructure in the next chapter.
The Supple Organization
Taken together, a transparent change culture and flexible organizational structure play an important role in shaping employee perceptions about what an organization should be. The prevailing idea that organizations should be stable, fixed structures with clear reporting lines is the product of a stable and predictable environment. Facing a much more dynamic and uncertain landscape, Alibaba has prepared itself to evolve quickly by making change a part of business as usual for its employees and providing the clarity and infrastructure to support it.
Our company has thus gradually embraced the self-tuning ethos. We work to apply an evolutionary approach to all levels of the company. The vision, the business model, and even our organizational structure are regularly recalibrated to the environment through experimentation. Most fundamentally, our learning processes do not occur in a top-down, deliberative chain. They are spread throughout the organization and are self-directed. The organization is no longer viewed as a means to amplify and cascade leadership’s intentions. Information, be it user input, environmental changes, or effective or ineffective responses, flows freely through the organization, and every actor can respond. With the vision articulated by leadership acting as its magnetic pole, the organization moves organically.
Taken together, a focus on exploration, experimentation, and iteration across the business mandate a reconceptualization of organization and management. The firm becomes a collective machine for continuous iteration, much like an enormous algorithm whose objective function is the needs of the customer. The machine engages with its environment, its partners, and its clients, obtaining rich feedback, which it uses to further oil its gears and run more effectively. Management doesn’t tell the machine what to do; it merely makes sure everything runs smoothly.
If managers no longer direct and control, and the organization is self-tuning, then what is the role of management in a smart business? What should the firm itself look like, and how should managers design their organizations? We turn to these topics in the next chapter.