Parallel Execution Models
Why are we talking about parallel execution models today? Well, according to Forbes, “a staggering 70% of digital transformations fail.” While there is not just one reason why digital transformation (DT) efforts might fail, one of the biggest has to do with the fact that an organization still needs to run during its transition. In short, CEOs and CIOs alike recognize the need to digitally transform but often struggle with advancing a digital positioning initiative while simultaneously keeping the organization running. While there are many potential approaches, depending on how the organization is situated and what its goals are, too few organizations successfully balance the two. In other cases, leadership underestimates or misunderstands the effects that a digital business initiative has on the status quo. In either case, poor planning often leads to inappropriate support of the digital initiative; or worse, lack of coordination can undermine the existing business model and sink the business.
To help resolve this paradox, this article describes four parallel execution models that businesses, as well as government agencies, can use as a starting point. We will also detail here how to maintain a steady state while executing a transformation, starting with what digital transformation and positioning are and the triggers that may affect our approach.
What is Digital Transformation?
To some extent, DT is in the eye of the beholder, and what constitutes a digital transformation is less important than whether an organization benefits from an improvement initiative. For the uninitiated, let’s keep it simple:
Digital Transformation is when an organization uses technology to create new business models or ways of operating, which often result in new or different relationships with customers, new products, new services, or new combinations thereof, and which is usually dependent upon extreme efficiency and operational excellence.
Okay. The “simple” part may be a lie . . . but you get the gist. Check out this previous blog post for a deeper understanding of how IT can enable digital strategy, what digital transformation is, and what DT is not.
What is Digital Positioning?
Digital Positioning relates to specific strategies an organization employs to use digital transformation to gain competitive advantage. Depending on who you ask, there are generally two to four different digital positions an organization may attempt to achieve:
- Customer Excellence or Demand Shaping
- Operational Excellence
- Industry Transformation
- Market Transformation
As discussed in “Crafting Digital Strategy with the Digital Positioning Model,” organizations can use Beyond20’s digital positioning model (DPM), authored by yours truly, to determine the most appropriate digital position to take.
What are the Triggers of Digital Transformation?
Excepting the rare organization that transforms just for fun, the triggers that lead organizations to consider digital positioning are many and varied. On one end of the spectrum, some organizations take drastic defensive actions to counter sudden and unexpected threats from competitors and new technology. At the other extreme, some organizations make aggressive maneuvers to disrupt their industries or transform entire markets. To a large extent, some combination of intentional digital positioning and DT triggers inform how to maintain a steady state while transforming.
What is a Parallel Execution Model?
Parallel Execution Models (PEMs) are approaches to executing digital strategy while maintaining the “steady state” of the business. In other words, with the exception of start-ups and organizations starting entirely new subsidiary brands, most organizations need to keep the business running while changing the way they do business. PEMs also help to cope with what Wired calls the “Innovator’s Dilemma,” where one product or service starts to become popular and disrupts the existing marketplace; in turn, this new offering itself becomes a target for disruption.
PEMs share a lot in common with both product portfolio strategies and the “4 Ps,” namely “product,” “price,” “place,” and “promotion” of the marketing mix. When companies determine which products and/or services to offer, they realize that successfully promoting certain offerings may have a negative effect on other offerings; whereas in some cases it is possible to create complementary offerings. Likewise, in marketing, consideration of two of the Ps, “place” and “promotion,” raises similar questions. For example, selling a particular product or service to one customer may mean the customer neglects another offering. In this case, selling a product or service at a corporate location may have the effect of stealing sales away from franchises and re-sellers.
Business Model – Description of means and methods a firm employs to earn the revenue projected in its plans. It views the business as a system and answers the question, “How are we going to make money to survive and grow?” (BusinessDictionary.com)
However, PEMs are broader than product, place, and promotion (though they often involve consideration of all of these factors). PEMs address the business model itself. Using terminology borrowed from the field of evolutionary biology, PEMs can be reduced to four basic models:
- Opportunistic Cannibalism
- Parasitic Mutualism
- Symbiotic Synergism
Before proceeding to describe each, it should be noted that rarely do any of these PEMs exist in a pure state. It is also entirely possible to graduate or transition from one PEM to another.
Parallel Execution Model #1: Opportunistic Cannibalism
Cannibalism is defined as an “ecological interaction” that involves “the act of consuming another individual of the same species as food.” – quite the “interaction,” for sure. In the animal and plant kingdoms, it may well be considered the most aggressive strategy for survival and is often an extreme reaction to threats. It is often driven by over-crowded spaces with many competitors and dwindling sources of nourishment.
In the world of DT, opportunistic cannibalism occurs for many of the same reasons. An organization realizes (often quite suddenly) that its business model (and possibly its products and services) is no longer viable or under serious threat from competitors or radically new applications of technology. Similar to the animal kingdom, when too many competitors are in the same market or too many similar or substitute products are being sold to a customer base, an organization may be willing to take extreme measures.
Pages Jaune (the French Yellow Pages) serves as a prime example of opportunistic cannibalism. Around 2009, Pages Jaune was the market leader of the French Yellow Pages industry. The only problem was that the industry was crumbling . . . the fact was that nobody wanted to thumb through a thick paper phone book when they could find even better information online. As consumers started to use printed Yellow Pages to prop open windows on warm Paris Junes, Pages Jaune print revenues plummeted by more than 10% year over year. The new CEO, Jean-Pierre Remy initially struggled to sell a digital transformation to employees. After all, Pages Jaune was the industry leader, had been successful for most of its history, and had survived other technological and industry shake-ups. Mr. Remy understood that printing paper directories was not Pages Jaune’s forte. Instead, their strength was the ability to connect businesses to local consumers.
Remy boldly and unequivocally declared the old business model of printed books dead. His goal was to have digital revenue account for more than 75% within five years (at the time, it accounted for less than 30%). He also put a freeze on investment in the print directory business model (unless absolutely necessary for survival) during the transition. Remy managed to pull off his great tour de force within a little over four years as digital revenues replaced the losses from the print side of the business. In 2015, Pages Jaunes reported revenue growth for the first time in years.
Intentional, Radical Destruction and the “Engine of Cannibalization”
In the case of Pages Jaune, transformation and cannibalization of the old business model was triggered by the existential threat posed by drastically declining revenue. Not all organizations that choose opportunistic cannibalization are forced into it. For some organizations, cannibalization is more proactive – It is the intentional and radical destruction of the old. For example, Netflix has transformed several times, from a rental DVD-to-consumer provider to a streaming video provider to a content creator and has become a veritable “engine of cannibalization.” Not only did it cannibalize its own model of DVD-to-consumer in favor of streaming, it also intentionally cannibalizes its own content. As one Netflix-original program begins to peak in popularity, Netflix intentionally creates and promotes new content. Although this may have the effect of drawing some customers away from existing content, it also helps Netflix grow the consumer base.
If you don’t cannibalize yourself, someone else will.
– Steve Jobs
A third leading example of constant cannibalization is Apple. They have intentionally cannibalized the Mac with the iPad, the iPod with the iPhone, and the iPhone with the . Interestingly, in several of these cases, cannibalization turned into complementary products and synergy since many consumers own more than one apple product, including products that provide some crossover functionality. As Steve Jobs aptly said, “If you don’t cannibalize yourself, someone else will.”
Parallel Execution Model #2: Parasitic Mutualism
In nature, parasitism occurs when “one organism lives on another organism, is adapted to its way of life, and causes harm to the host.” For a relationship to be strictly parasitic, three conditions must be present, 1) The host can survive without the parasite, 2) The parasite cannot survive without the host, and 3) The host is harmed by the parasite. For example, the roundworm needs the human host to survive; but the human would rather not nurture the roundworm. In fiction, Dracula’s vampirism is purely parasitical.
From a DT perspective, parasitic mutualism can be considered a “kinder, gentler” form of cannibalism. To be sure, the intent is to benefit from the still-lucrative resources of the existing business model for as long as possible. If the existing business model is financially successful, profits and other resources can be used to fund the emerging model. Arguably, almost all PEMs involve some form of parasitic mutualism since new digital models need to be funded and the old business models provide revenue to maintain the business as a going concern.
All metaphors break down at some point. In the natural world, the strategy of the parasite is to preserve the host for as long as possible, even indefinitely. Although incremental harm is done to the host, the parasite incrementally benefits. The opposite is likewise true – Destroying the host means the end of the meal ticket. To be clear, in DT, the intent is to eventually destroy the existing business model in favor of the new. It can be thought of as “slow decay” or “measured destruction.”
For example, the New York Times (NYT), a leader in the newspaper world, was threatened by declining revenue from print-based advertising, which represented a significant revenue stream for print news publications. The NYT predicament was not unique to their company and was distressing to the entire industry. Going against conventional wisdom at the time, NYT decided not only to provide online news content but also to charge for it. This represented a bold experiment and one that carried significant risks. First, establishing the new digital business was not free. In fact, NYT used profits from its existing print business to subsidize the digital business. This meant that for several years, NYT’s overall cost increased since they were effectively running two businesses in parallel – the old print business and the new digital business. Second, although NYT was sure that “going digital” was predicted by consumer demand, it was not clear if or when the old print business would go away. Thus, a radical cannibalization was less apparent than a slower (though still aggressive) transition. All’s well that ends well for NYT – by the end of 2017, print and digital subscription revenue combined was about double its print-based advertising revenue.
The New York Times is a classic example of where bold action and clear direction successfully employed parasitic mutualism in a proactive sense. However, The New York Times consciously chose this approach. To the contrary, for too many organizations, parasitic mutualism becomes the default option when leaders cannot make up their minds regarding and, unfortunately, do not end up with the same level of success (nor come remotely close) to the success of the NYT. As usual, no plan is not much of a plan.
For too many organizations, parasitic mutualism becomes the default option when leaders cannot make up their minds regarding what approach to take.
Parallel Execution Model #3: Commensalism
Commensalism is a form of symbiosis in which one organism gains benefits from another organism while the “host” neither benefits nor is harmed. The remora fish is the ocean’s great commensalist. It has adapted its body and habits to attach to larger creatures such as sharks . In addition to hitching a ride on larger, faster fish, the remora benefits by eating the larger creature’s fecal matter. (Nobody ever said nature was pretty.) Although the shark gains no obvious benefits from the remora, it also is not harmed by it. Luckily for the remora, the shark has no interest in eating it.
Commensalism in DT works best when the organization is attempting to gain market share or get a larger slice of the market pie. It also works well when certain consumers are hard to reach through existing channels. For example, a local haberdasher does good business with local clients who shop in-store for hard-to-find clothing accessories. The haberdasher knows that the geographically local market is loyal but relatively small. They also recognize that potential consumers exist outside the local geography and are not able to travel to visit the store in-person. Therefore, the haberdasher opens an online shop where they sell a small subset of their goods. Thus, the online store builds on the haberdasher’s brand and existing business model without compromising it.
It is noteworthy in this case that there is no “place-based” competition between the physical and online stores. Local customers continue to visit the physical shop (and may even prefer the “personal touch” and “old world feel” they get there); online customers only (or primarily) visit the online store. Additionally, note that there is no major diversion of resources from the old model to the new model. In fact, both models peacefully co-exist. Each model could exist without the other.
One could argue that in the scenario presented here, the online store does not truly represent a digital transformation and that it is simply an additional sales channel. This is a fair point; and for this reason, commensalism is often used in channel strategy positioning.
In the biological sciences, some argue that there is no true commensalism. If the remora in some minor way slows down the shark, then isn’t parasitism at work? If the remora removes potentially dangerous bacteria from the shark, isn’t that mutualism? Don’t some of the profits from the physical haberdashery go to support the online store? Technically, all of these can be true . . . but again, organizations are more interested in successful outcomes than academic precision. Which brings us to our fourth and final approach, Symbiotic Synergism.
Parallel Execution Model #4: Symbiotic Synergism
In symbiotic relationships, organisms live in a close nutritional relationship and rarely live outside of the relationship. In synergistic relationships, two or more organisms act together to produce a substance that none can produce separately. In other words, two organisms work better together than apart. For example, the enjoy a synergistic relationship. When they work together, both are better able to defend themselves from predators.
The anemone stings fish that prey on clownfish. In turn, the clown fish scares away butterfly fish that eat anemones. The hummingbird and the flower likewise share a symbiotic synergistic relationship. The hummingbird relies on the flower for nectar. The flower relies on the hummingbird for pollination.
The basic idea behind symbiotic synergism is that two things combined produce a greater or different result than they could have produced individually. Like peanut butter and chocolate, some things just go better together: 1+1>2.
Commensalism is often used to gain a larger piece of the pie and expand market share, whereas symbiotic synergism can be used to grow the market itself and effectively expand the pie. Symbiotic synergism can be ideal in situations where adjacent or complementary sales channels exist, there is no competition amongst channels, and customers have a preference for multi- or omni-channel sales. For example, this includes customers who want to shop in a physical store sometimes and online at other times.
Sephora, the cosmetics retailer, is a textbook example of this approach. Although they recognize that a critical aspect of selling new make-up is the ability to try it on in a physical store, they also learned that in-store customers are also highly influenced by online and video reviews and recommendations from social media influencers. As part of the design of their mobile application, they allow in-store consumers to scan codes from physical displays which navigates the customer to videos, tutorials, and product reviews. Once a customer becomes familiar with a product, they may decide to re-order the same product from the online store instead of traveling to a physical location. Additionally, for customers regardless of where they are shopping, Sephora created a Virtual Assistant application that allows customers to virtually try on various make-up products. Thus, the physical store business model supports the online business model and vice versa, all aided by digital technology like the Sephora mobile application, Virtual Assistant, and geolocation technology.
A critical aspect of the Sephora business model is that physical stores are not franchised. Sephora is owned by LVMH (the same company that owns Louis-Vuitton, Moet, Hennessy, Tiffany, Christian Dior, and in total about 75 brands). LVMH operates more than 2,600 Sephora locations worldwide, which gives it complete control in terms of the in-store experience and how it interacts with the virtual world. It also means that there is no competition between owners of physical stores and the online store. If that were the case, Sephora would have to re-think how the digital business model could co-exist with the physical store model without cannibalizing it.
The Symbiotic Synergism model also tends to work well when complementary products exist. For example, customers who buy an iPhone are likely to want to listen to music on their phones, which creates an opportunity to sell downloadable music through iTunes. In turn, this creates an “indirect network effect,” where a customer’s use of one product reinforces the use of complementary products. Additionally, the use of complementary products potentially increases customer “switching costs,” which has the result of increasing customer brand loyalty.
Given the focus on sales channels and complementary products, organizations often use some version of symbiotic synergism to achieve improvements in the customer excellence digital position. Real masters not only meet customers where they are by providing omni-channel delivery, they also engage in demand shaping. In other words, these organizations go beyond reactively responding to consumer demand. Instead, they proactively create demand through targeted channel marketing and even individual consumer promotions.
Zone of Confluence and Suggested Practices
Although it is not appropriate for every organization, in many ways, symbiotic synergism is the most mature or advanced of the PEMs. It may provide the greatest benefits but is also the most difficult to execute. In large part, this is due to the relationship between the “customer excellence” and “operational excellence” digital positions. It is difficult to earn lasting improvements in customer excellence if tremendous operational efficiencies do not also exist. Just think of the highly-digital transformer, Amazon. Satisfying customers by rapid delivery is not possible unless the operational and fulfillment back-end and logistics are optimally efficient. For this reason, organizations interested in symbiotic synergism would be wise to consider defining their unique zone of confluence practices and focusing on maturing practices that help improve both operational efficiency and the consumer experience. Every organization may identify different practices within the zone of confluence.
However, practices that tend to give organizations a “two-for” in terms of improving more than one digital position include the following practices:
- Strategy Management,
- Relationship Management,
- Financial Management,
- Workforce and Talent Management,
- Data Analytics,
- Measurement and Reporting,
- Organizational and Change Management,
- Infrastructure and Platform Management, and
- Product or Program/Project Management
Five Characteristics of Fragile Models
The worst PEM is none at all. Organizations that follow the “no model” approach are either extraordinarily lucky or are no longer in business. Too often, organizations eventually muddle into a model due to fear, procrastination, and ignorance. The result of this approach (or lack thereof) is that the organization has few options for how to execute on digital strategy. Other organizations create or adhere to a PEM, but the model is largely inappropriate to meet the needs of the organization. Generally speaking, fragile models are either poorly defined or too rigid to be practical. Characteristics of fragile models include the following:
- Constructed without an outcomes-based view
- Do not account for alternate models or realistic scenarios
- Constructed as a response to extreme risk attitudes
- Based on a misunderstanding of the current business model
- Are not sufficiently flexible and do not account for the potential need for repositioning
The best way to avoid fragile modeling is to assemble a small dedicated team when crafting PEMs. While the team should be small enough to work quickly and make decisions, it should include representation from the larger organization, especially those familiar with organization-critical risks and opportunities. In short, it is less about how many people are on the team and more about who is on it.
“Perishability” and Rational, Evolutionary Modeling
Designing PEMs with the notion of “perishability” in mind is another way to defend against a fragile model. Perishability is the philosophy that something that is created has limited use and will at some point no longer be useful. It can be helpful to think of a PEM (or any model, for that matter) in this way. At times, organizations become so tied to a model that they doggedly pursue it well beyond its usefulness.
A better approach is to consider a Rational Evolutionary Modeling (REM) approach. The REM approach accepts upfront that the optimal model is not immediately achievable and substitutes an inferior or minimally viable model in its place. The longer-term goal, however, is to work towards replacing the temporary model and transitioning to the next model when certain environmental conditions are met or trigger events materialize. For example, an organization may ultimately wish to cannibalize its existing business model but recognizes that it is too dependent on existing revenue streams or does not possess the current capabilities to make a radical change. Instead, the organization may opt to employ parasitic mutualism as a short-term solution with the goal of adopting a cannibalistic approach as soon as possible. Likewise, an organization may wish to use a synergistic approach but lacks the digital technology to enable multichannel symbiosis. It decides to follow a commensalist approach with the notion of improving digital technology maturity until it can evolve to a synergistic approach.
Pace of Transition
Just as the best PEM varies per organization, so too does the pace of transition. When executives are asked about how quickly change should be implemented, typically two extreme schools of thought emerge: The first school suggests a rapid (albeit possibly painful) transition period. The second suggests a longer-term approach due to the uncertainties inherent in “going all in.” There are no easy answers. What is clear is that four factors need to be considered:
- Consumer Demand
- Organizational Capabilities
- Maturity of supporting Digital Technologies
- Threats from Competitors and Emerging Technologies
A good axiom for any organization to consider is “when in doubt, ask the consumer.” In some cases, consumer demand dictates that the business model change sooner rather than later. Take the life and health insurance business in the United States as an example. At some point half a generation ago, consumers felt more comfortable purchasing from a real human being insurance agent whom they trusted. These days, consumers would rather purchase anonymously online without ever having to spend time with a person. This is potentially a win-win for the savvy insurance company . . . the consumer buys from the channel they prefer, and the insurance company can reduce costs since it needs fewer agents. To be clear, consumer demand is often the number one driver of transformation pace.
An organization’s ability to execute on a digital transformation and to control parallel business models is possibly the second biggest driver of pace. There is considerable debate in this area. One school of thought suggests that an organization should ignore current strengths and weaknesses (which support old ways of working) and instead aggressively plow forward by hiring people with new skills needed to support the “to be” or “future state” business model and expedite the purchase of new technologies. A more tempered approach suggests taking stock of capabilities the organization currently has, understanding where the organization wants to be, and performing a gap analysis to understand what capabilities need to be built or acquired to bridge the gap. This is where using a tool like the digital positioning model comes in handy to help an organization understand how improving specific practices can help achieve a desired position. One of our customers, a major energy supplier, took this approach. We performed an assessment to understand the state of key IT practices like incident management, change management, and problem management as well as critical business practices such as relationship management and financial management. After performing the assessment, the energy provider understood that the relatively immature state of IT and business practices as well as low maturity with digital technology meant that transformation would likely be slow but needed to begin immediately.
Maturity of Supporting Digital Technologies
Technology in and of itself is normally not the only solution to a problem. Having said that, technology puts the “digital” in DT, and it is hard to imagine a transformation without significantly using technology. Although there is some debate whether digitalization is really just extreme automation in disguise (I think it’s much more than automation), a significant amount of automation is usually a prerequisite for launching a digital transformation. For example, automation can free-up employee time to perform higher-value tasks such as relationship building, complex problem solving, and consumer demand shaping. With automation as a starting point, organizations should consider what digital technologies would best support their “to be” digital business model. Generally speaking, the Legal industry is a mature industry that has established business models, but has not made great use of digital technology. Some have even suggested that due to the maturity and complacency of the industry that it is ripe for disruption from a competitor in a completely different industry. However, there are some emerging technologies to support various aspects of the industry. For example, technology exists that allows customers to compare legal documents to law firm documents in a database, which minimizes the need for the law firm to pay dozens of paralegals and junior lawyers and reduces costs for clients. Technology to help improve internal decision-making also exists. For instance, software can look at certain legal cases and using a combination of rules-based and fuzzy logic to determine the probability of winning a case and the anticipated award or return on investment.
25 Digital Technologies to Know
Artificial Intelligence (AI)
AI-Enabled Event Management
Integrated Service Desk
Augmented and Virtual Reality
Natural Language Processing
Big Data and Data Analytics
Robotic Process Automation
Mobile Payment Systems
Moment-Based Marketing Applications
Customer Relationship Management Systems
Threats from Competitors and Emerging Technologies
Needless to say, it is important to know what competitors in the same industry are doing. There may be times when a competitor’s adoption of technology warrants an organization to accelerate its rate of change. At the same time, a wise organization avoids rushing into adopting a particular digital business model or technology to simply “keep up with the Joneses.” A model of technology that works for one organization may not work for another. In some cases, it is more prudent to closely monitor a competitor to determine whether its digital business model or technology adoption is paying off before investing in a losing endeavor. Although considerably more difficult, it is also important to monitor well-funded digital savants in other industries. It may be impossible to predict what markets the Amazons, Googles, and Facebooks of the world will move into next. It is even harder to counter the threat even if it can be predicted. At the same time, not all of the big players have been successful with every opportunity (read on to learn about some of Amazon’s failures); and this reality presents an opportunity for smaller organizations to carve out niche offerings.
In certain cases, a technology (and sometimes even a particular technology vendor) has been so successful in marketing potential benefits that a frenzy of organizations rushes to beat its competitors in adopting it; though no organization fully benefits. For example, a recent article in Beckers Hospital Review cites a New England Journal of Medicine claim that despite widespread adoption of Electronic Health Records systems, healthcare has fundamentally failed to innovate. Electronic Health Records (EHR) Technology promised to transform the patient experience (make it more convenient and safer), improve patient health outcomes, and improve the healthcare provider’s operational efficiencies. Arguably, electronic health records systems have not fully achieved any of these outcomes despite significant provider investment. This may be largely due to healthcare providers’ willingness to invest in new technologies without a corresponding appetite for changing business models. This is not to say that healthcare providers should not invest in EHR (there are many compliance- and safety-based reasons to do so). It does suggest, however, that blind adoption of technology without thought for the business model introduces significant risks.
Continual Improvement is the New Black
One thing in common to all organizations that have mastered digital transitions is their appetite for change. Far from becoming complacent with success (whether attributed to old business models or digital models), these organizations are never satisfied with even industry-leading results. For them, continual improvement is the new black.
There’s nothing about our model that can’t be copied over time. But you know, McDonald’s got copied. And it’s still built a huge, multibillion-dollar company.
– Jeff Bezos
Once again, consider Amazon. They were founded in 1994 as a physical goods company selling print books online. Within just two months of doing business as an online merchant, they were selling books across the United States and in more than forty countries. Their weekly sales were already $20,000. This initial success would have satisfied many narrow thinkers. But founder Jeff Bezos was thinking bigger. Along the way, Amazon expanded to selling products other than books (e.g., music CDs), which is a logical but still narrow form of growth. Despite success with selling print books, Amazon introduced the first Kindle electronic reader in 2007, intentionally and slowly cannibalizing their successful print book business (along with the print book business at large). In a larger move, over time, they shifted their business model from being a product-based company to a platform company that tolerates and even invites competitors to sell on their site. To be fair, Amazon does not have one business model nor is it one kind of company. In fact, many of Amazon’s business models are replicable by competition. In an interview, Bezos once said, “There’s nothing about our model that can’t be copied over time. But you know, McDonald’s got copied. And it’s still built a huge, multibillion-dollar company.” Today, they are involved in e-commerce, cloud computing, online streaming, artificial intelligence, electronics, government contracting, and grocery stores.
If the size of your failures isn’t growing, you’re not going to be inventing at a size that can actually move the needle.
– Jeff Bezos
Of course, there were failures along the way. For instance, in 1999, Amazon made a feeble attempt at book publishing but abandoned the attempt when it appeared it would be unsuccessful. In 2014, it introduced the Fire phone, which likewise failed to the tune of a $170 million loss. It launched Destinations, a travel site that failed even quicker than the Fire phone. Amazon local (think Groupon), Amazon wallet, local register, TestDrive, and WebPay were all likewise failures. The key to Amazon’s success is the willingness to take risks and learn from failure. According to Bezos, “If the size of your failures isn’t growing, you’re not going to be inventing at a size than can actually move the needle” and “The good news for shareholders is that a single big winning bet can more than cover the cost of many loses.”