What is incompatible innovation and why is it as dangerous as disruption?

Corporations are fine-tuned to doing business one way. This sends them struggling when someone comes up with a better way.

Christian Vogt
Jun 27 · 21 min read

Anyone listening to the talks at even the most reputable corporate-innovation conferences would be forgiven for thinking that the term “disruptive innovation” refers to any innovation with material impact. Uber, iTunes, Tesla — they all sound quite disruptive, do they not? But a closer look at the work of Clayton Christensen, who developed the theory of disruption and coined the term, reveals that disruptive innovation is something more specific: It is a theory of market entry and competitive response that predicts that, when innovation enters a market from the low end or a niche, it is commonly ignored by the incumbents, whose priorities are the higher-value market segments. If subsequently the innovation improves on a steeper trajectory than established alternatives do, say, thanks to a new technology it builds upon, it can eventually expand into the higher-value market segments that the incumbents covet. It is then often too late for the incumbents to catch up because they, too, have to follow the learning curve the new market entrant has gone through. At this point, disruption is said to have occurred. Neither Uber nor iTunes nor Tesla were disruptive innovations according to this theory, since all of them attacked either the mainstream market (as in the case of Uber and iTunes) or a high-value market segment (as in the case of Tesla) from the start. Using the terminology of Clayton Christensen, Uber, iTunes, and Tesla were instead sustaining innovations.

Of course, Uber, iTunes, and Tesla were without doubt breakthrough innovations that dislodged market-leading incumbents and transformed industries. They may not have been disruptive, but they were tremendously impactful nonetheless. What is more, incumbents struggled to take advantage of these innovations, a phenomenon similar to what happens in the case of disruption. Is there something that these innovations have in common, beyond an amazing idea and impeccable execution, that made them so successful?

It turns out there is a common pattern behind the success of Uber, iTunes, and Tesla — as well as an array of other non-disruptive innovations — that allowed the new market entrants to leave the incumbents in the dust. In all cases, the new market entrants leveraged a novel business model, often enabled by a new technology, that allowed them to deliver an existing industry’s value proposition in a significantly better way, but with organizational structures or processes that were materially different from those used by the incumbents. As a result, the incumbents were unable to copy the new market entrants, even though they were acutely aware of the superiority of the entrants’ approaches.

The common pattern warrants giving this type of innovation its own name: incompatible innovation. This article examines how incompatible innovation works, how it is different from disruptive innovation, and why it is just as dangerous to incumbents as disruptive innovation. In the process, we will also understand why incumbents always need to create an independent entity when taking on incompatible innovation, whereas they do not necessarily have to in the case of disruptive innovation. Now, to understand incompatible innovation and the reason it is dangerous, we first need to take a step back and look at how corporations function, what makes them successful, and what it is that limits their ability to respond to certain types of innovation.

All corporations were built to execute one or a set of specific business models. They execute these business models well because their organizational structures and processes have been optimized for them over years, often decades, of continuous improvement. Yet for the same reason, corporations are also slow to experiment with new ways of executing these same business models. Experiments require exceptions to organizational structures and processes, and hence get in the way of “business as usual.” The exceptions take a herculean effort to uphold because everyone involved is still compensated on — and more familiar and comfortable with — the old ways of working. This makes corporations slow to adopt a new and better way of doing business when a novel business model emerges that, perhaps on the back of a new technology, enables this. Execution is the strength of all corporations since this is what they were built for. Experimentation, by implication, is not.

Early-stage startups are the opposite of corporations in this regard. Because startups are still in search of a viable business model, their strength is experimentation. Only when they have found a business model that works and begin focusing their attention on scaling will they build the organizational structures and processes to industrialize this business model, and in consequence gradually shift from being experimentation-focused to being execution-focused. On the spectrum between experimentation on the left and execution on the right, startups are all the way to the left, corporations are all the way to the right. Startups eventually shift right as they scale. No one ever shifts left.

Execution is the strength of all corporations since this is what they were built for. Experimentation, by implication, is not.

Organizational structures and processes are hence a two-sided sword. On one hand, they industrialize a corporation’s way of doing business and minimize its dependence on key individuals, thereby reducing risk and variability. On the other hand, they limit the corporation’s ability to experiment and respond to innovation. Because organizational structures and processes are so determining for a corporation’s effectiveness in handling innovation, let us look more closely at their merits and limitations.

Clayton Christensen observes that organizational structures and processes become a key driver of value for a company as it evolves from early-stage startup into corporation. It is widely acknowledged in venture capital that the value of an early-stage startup rests almost entirely on its team — its experience, camaraderie, and grit. Early-stage startups exist to turn an idea into a business model that works and scales, and you need stellar talent to do this. There are few organizational structures and processes at this early stage because the precise business model to which they would cater is yet to be found. Organizational structures and processes would only get in the way of the experiments the team needs to conduct to identify and refine the business model it wants to develop.

However, once the right business model emerges, the startup will start building organizational structures and processes around it. The quality of these structures and processes determines the robustness and repeatability by which the startup makes money, and it reduces the startup’s dependence on key individuals. If a startup fails to establish effective organizational structures and processes, its success will continue to hinge on its team. Such a startup is at risk of faltering once key people leave, as is often the case after an exit. Establishing effective and efficient organizational structures and processes is therefore crucial for a startup to scale into a corporation.

On the spectrum between experimentation on the left and execution on the right, startups are all the way to the left, corporations are all the way to the right. Startups eventually shift right as they scale. No one ever shifts left.

But organizational structures and processes not only define what a corporation can do, they also define what the corporation cannot do, and even what it does not want to do. There are two ways in which this happens, one limiting a corporation’s ability to react to incompatible innovation, the other limiting its ability to react to disruptive innovation. First, the organizational structures and processes themselves dictate what a corporation can do and cannot do. They prevent the corporation from embracing incompatible innovation because incompatible innovation, by definition, is at odds with how the corporation is used to doing business. Second, the organizational structures and processes indirectly keep the corporation from embracing disruptive innovation because they shape the priorities based on which employees make decisions and allocate resources. This will cause the corporation to deprioritize disruptive innovation for one or more of three reasons:

  • Organizational structures and processes are geared towards a certain customer segment, so the needs of this customer segment will get priority over the needs of other segments. Disruptive innovation initially targets a new customer segment — the low end of the market or a niche — and will therefore not be considered a priority.

The organizational structures and processes of a corporation hence directly determine if it is feasible for the corporation to adopt a given innovation, and they indirectly determine, by shaping the priorities of the corporation, whether an innovation is deemed desirable. They render all incompatible innovation infeasible, all disruptive innovation undesirable: Incompatible innovation is innovation that the corporation wants to adopt, but cannot, because its existing organizational structures and processes make this infeasible. Disruptive innovation is innovation that the corporation can adopt, but does not want to, because its priorities for decision-making and resource allocation make it undesirable.

The corollary of this line of reasoning is that incompatible innovation is not the same as disruptive innovation, and compatible innovation not the same as sustaining innovation. Whether an innovation is compatible or incompatible depends solely on its fit with the organizational structures and processes of the corporation. Can the corporation adopt the innovation and still largely continue to do business as it used to? Or would adoption of the innovation require significant alterations to the organizational structures and processes of the corporation? On the other hand, whether an innovation is sustaining or disruptive depends on the market segment it is initially addressing. If the initial market segment is a segment that incumbents have little interested in, that is, the low-end of the existing market or a niche, and the plan is to use this initial market segment as a bridgehead into the mainstream market, then the innovation is disruptive. If the innovation addresses the mainstream market straight away, it is a sustaining innovation.

A corporation’s organizational structures and processes render all incompatible innovation infeasible, all disruptive innovation undesirable.

With this groundwork, let us look again at the three examples from the beginning of this article: Uber, iTunes, and Tesla. While none of them was disruptive innovation, we now know that the reason for the success of each of them was the fact that they were all incompatible innovations.

Uber — Uber, along with its competitors like Lyft and Didi Chuxing, have undoubtedly transformed the taxi industry and redefined mobility. But Uber has not been disruptive because it did not enter its market from the low end or via a niche. It squarely attacked the mainstream segment of the taxi industry with a service that was on par with, if not better than, existing alternatives. However, Uber’s mobile-first, asset-light business model was decidedly incompatible with that of established taxi companies. This left the taxi industry unable to come out with a similar experience at a comparable scale.

One caveat: There has been some anticipation that ride-sharing will reduce car ownership. Whether this happens remains to be seen as, for now, car ownership in the United States is still on the rise. But there appear to be niches, especially in urban environments, in which people do ditch their cars in favor of ride-sharing. If this nascent behavior turns into a more material trend over time, it would indeed be a case of disruption. To car ownership, however; not to the taxi industry.

Apple iTunes — Apple’s iTunes did not enter the music industry from the low end or via a niche. But its online, unbundled, by-the-song business model was decidedly different than the sale of albums, largely in brick-and-mortar record stores, that was the bread and butter of the rest of the industry. iTunes was therefore not disruptive to, but incompatible with, the business model of the incumbents of the music industry at the time. Similarly, the next wave of innovation in the music industry, subscription-based streaming, is not disruptive either, but again incompatible, this time with the unbundled sale of individual songs.

Tesla — Tesla was not disruptive because its original sweet spot was the luxury segment of the auto market, the industry’s most valuable segment. A disruptive approach, in contrast, would have been to build an electric car that was cheaper than existing alternatives (addressing the low end of the market) or to build an electric car with a set of properties otherwise unavailable in the market (addressing a niche of the market). In China, some manufacturers of electric cars do both. Their cars are cheap and narrow, allowing merchants to deliver goods on tight roads. Tesla, though, did neither.

What allowed Tesla to move faster than incumbents was its direct-to-consumer distribution model, a model incompatible with the way distribution of cars had been done traditionally. Incumbent automakers, who sell through independent dealerships, struggled with electric cars primarily for three reasons: First, electric cars have longer sales cycles than combustion engine cars because they are a new technology and customers take longer to make a purchasing decision. This makes them less attractive to dealers, causing dealers to prioritize selling combustion engine cars over selling electric cars. Second, dealers generate nearly 50 percent of their gross profits from after-sales maintenance and repairs, but electric cars require notably less maintenance than combustion engine cars. This gives dealers another reason to sell electric cars with caution. Third, without their own nationwide location footprint, and lacking strong support from their dealerships, incumbent automakers failed to develop a credible charging infrastructure that customers could rely on. Tesla addressed all these challenges by integrating three elements of the value chain that were traditionally separate — manufacturing, dealership, and charging infrastructure (the element of the value chain traditionally served by gas stations) — a move that incumbent automakers were unable to copy. Granted, incumbent automakers have caught up. But it may not have happened if it had not been for strong government incentives, a wide-ranging scandal over exhaust emissions manipulation, and a rising concern about climate change.

Corporate innovators should take incompatible innovation serious as it is just as dangerous as disruptive innovation. In both cases, incumbents fail to take advantage of an innovation that can lead to the demise of their existing business. In the case of disruptive innovation, incumbents choose to not take advantage, even though they could. In the case of incompatible innovation, incumbents cannot take advantage of the innovation, even though they might see the need.

Of course, Uber, iTunes, and Tesla are not the only examples of innovation that was incompatible, but not disruptive when it hit the market. The list goes on. Here are a few more examples:

Amazon — Amazon, and e-commerce in general, was not a disruptive innovation as it directly attacked the mainstream retail market. More specifically, the online sale of books, the sub-segment of e-commerce where Amazon started, was not disruptive either for the same reason. But Amazon was an incompatible innovation that did away with brick-and-mortar outlets and instead required new competencies like an intuitive online-discovery experience, artificial-intelligence-based product recommendations, seamless checkout processes, highly automated warehousing, fast and cost-effective last-mile delivery, and impeccable remote customer service. Traditional retailers did not have these capabilities.

Apple iPhone — Apple’s iPhone, which is sold at a premium in the smartphone market and has, for the most part since its introduction, only increased prices, is not disruptive. But Apple’s organizational structures and processes for developing the iPhone are incompatible with those of its competitors and were key to developing the user-friendly interface that made the iPhone popular. They were so unique, in fact, that all then-prevailing mobile operating systems have essentially vanished from the market, and today virtually all sold smartphones are either iPhones or smartphones based on the Android operating system. Moreover, the App Store, which Apple launched about a year after the first iPhone came out, was an incompatible innovation as well. Third-party mobile applications quickly became one of the most popular purchasing criteria for the iPhone. Behind the lineup of mobile applications was a carefully nurtured ecosystem of developers and an infrastructure of open programming interfaces that only Apple had at the time.

One caveat: While the iPhone was not disruptive to the smartphone industry, it may be argued that it was disruptive to personal computers. The broad availability of mobile applications, combined with the ease of use of the phone without a keyboard or a mouse, may have turned the iPhone into a supplement for the personal computer for some users.

Online travel agencies — Similar to how Amazon changed the retail industry, online travel agencies changed the travel industry. Until the 1990’s, flights, hotels, or cruises used to be booked primarily over the phone or in person with travel agents. Online travel agencies changed that. Their use of online distribution combined with a self-serve customer interface was incompatible with the way incumbents did business at the time. But because they did not start at the low end of the market or in a niche, and instead attacked the travel market right in its mainstream segment, online travel agencies were not disruptive.

Word processors — When word processors began to displace typewriters, they could not have been more incompatible with the way the typewriter industry functioned. The key competence changed from precision manufacturing to software engineering, the channel changed from servicing dealers to computer resellers, and the aftermarket moved from repair and maintenance to occasional software updates. But word processors were not disruptive. They attacked the mainstream typewriter market heads-on.

As the examples above illustrate, there is a new business model behind every incompatible innovation. It is this new business model that calls for organizational structures and processes different from those in use by the incumbents. The new business model, in turn, is often enabled by new technology. For example, Uber would not exist without the smartphone. The World Wide Web was the foundation for Amazon and online travel agencies. And word processors were possible only thanks to the personal computer, a new technology at the time. But not every incompatible innovation builds on a major technological breakthrough. For example, iTunes’ only major technical advantage over existing portable digital music players at the time was its FireWire technology that allowed iPods to be synchronized faster. Tesla uses commodity lithium-ion cells for their cars. And the original iPhone, even though some of its components may have been more refined than those built into other contemporary smartphones, was essentially made from the same categories of building blocks that the incumbents were using as well.

There is a new business model behind every incompatible innovation. This is often enabled by new technology. But not every incompatible innovation builds on a major technological breakthrough.

The way disruptive innovation and incompatible innovation are defined does not make them mutually exclusive. There is an overlap between the two with innovations that are both disruptive and incompatible. These are the innovations that are most dangerous for incumbents. Incumbents will tend to ignore them because established priorities for decision-making and resource allocation make them undesirable. And even if an incumbent does choose to take such an innovation serious, that incumbent will find it infeasible to adopt the innovation because existing operational structures and processes do not fit. Here are a few examples from the intersection of disruptive and incompatible innovation:

Wikipedia — Wikipedia was a disruptive innovation because it entered the encyclopedia market from the low end. At the time, most encyclopedia customers expected high-quality articles written by paid experts. Wikipedia, which crowd-sourced articles from volunteer editors, did not appeal to them as a convincing alternative. Encyclopedia publishers, too, took no interest. Wikipedia was a free service funded only through donations, and for them there was not enough money to be made with this business model. Only over time did Wikipedia emerge as a credible substitute for even the most renown encyclopedias. As a result, most established encyclopedias — paper-based ones and, as in the case of Microsoft’s Encarta, even premium digital ones — have exited the market. Wikipedia was also an incompatible innovation. Its business model, which comprised crowd-sourcing for content creation, sophisticated algorithms for quality assurance, online distribution, and donation-based monetization, was entirely different from that of the established encyclopedia publishers.

AirBnB — AirBnB is a disruptive innovation because it addresses the low-end market of the hospitality industry. The apartments and houses offered on its platform are those of other users, not professionally managed hotels, so the quality and service that customers can expect on AirBnB is typically not on par with that of hotels. At the same time, AirBnB is an incompatible innovation. Its platform-based business model makes money off of matching supply and demand, not by renting out property on its own balance sheet, so its organizational structures and processes are different from those typical for the hospitality industry.

Flash memory — Flash memory is a medium for permanent data storage that uses transistors to hold information. It is a substitute for the classic hard-disk drives that use magnetic domains on spinning disks to hold information. Flash memory was a disruptive innovation as it initially focused on a niche: It was too costly to be attractive for the traditional applications of hard-disk drives such as data centers or personal computers. So its initial use cases centered around those where hard-disk drives were too large or too sensitive to shocks, like thumb drives, portable music players, or digital cameras. Only over time did prices of flash memory come down sufficiently to make the technology attractive for some of the main use cases of hard-disk drives. Today, most laptop computers use flash memory, data centers increasingly deploy it, and even cloud service providers have started to offer flash-memory-based data storage. Flash memory is also an incompatible innovation. Because it builds on transistors instead of spinning disks, its manufacturing process is more akin to that of computer chips than it is to the mechanical-engineering-centric process that is core to building hard-disk drives.

Netflix — Netflix was a disruptive innovation as it, too, started in a niche. Its original business model of delivering videos by mail was a great value proposition for some customers who valued the convenience of not having to visit a store to rent movies. But customers had to be willing to wait a day or two for the videos to arrive. And the selection Netflix offered was initially small and excluded new releases, partly due to the startup’s limited capital, partly due to Hollywood’s practice of releasing new movies in theaters first, then video stores, and only then television and streaming. Most customers wanted the latest releases, and they wanted to watch their videos right away, so Netflix was not for the mainstream market at first. As a result, Blockbuster, the leading owner of video stores at the time, decided to ignore Netflix despite repeated advances on the part of the startup.

But the uptake of broadband internet access allowed Netflix to fix the imperfections in its value proposition. When the company started streaming videos instead of shipping them by mail, customers were suddenly able to watch videos with less wait time compared to in-store rentals, not more, because they could now watch the videos instantly without even a store visit. The growing popularity of Netflix, in turn, convinced Hollywood to provide Netflix with new releases sooner. Suddenly, video availability on Netflix was better than with Blockbuster because not only were new releases now available, customers also did not need to worry about their favorite movies being sold out. Clearly, Netflix was also an incompatible innovation. Its easy-to-use website, video recommendation algorithms, online-delivery infrastructure, and ecosystem of partnering smart-television makers were assets that classic video store chains did not have. Conversely, Netflix neither needed a footprint of stores on its balance sheet nor deal with store operations.

Is there anything corporations can do to successfully respond to incompatible innovation? Fortunately, there is a way, but one way only: Corporations can tackle incompatible innovation by separating out its development and commercialization into an independent entity. All product-facing and customer-facing functions need to be decoupled from the corporation’s core organization to ensure that existing organizational structures and processes do not influence or inhibit the development of a new business model. At a minimum, these functions must include product management, engineering, marketing, sales, customer success, and customer support. Shared services such as legal, tax, compliance, and human resources may continue to be provided by the core organization. But even then should each shared service designate a member to serve as a liaison for the independent entity. This will ensure that the needs of the independent entity are addressed despite its initially small leverage relative to that of the core organization.

Incumbents always need to create an independent entity when taking on incompatible innovation, whereas they do not necessarily have to in the case of disruptive innovation.

In contemporary corporate-innovation discourse, the need to create an independent entity is often associated with disruptive innovation. But a closer look at any disruptive innovation for which an independent entity is indeed necessary will reveal that this innovation is also incompatible — it lies at the intersection of disruptive and incompatible innovation. The need for separation from the corporation is in this case a result of the innovation’s incompatibility with existing organizational structures and processes, not a result of its disruptiveness.

Creating an independent entity is mandatory for incompatible innovation, whether disruptive or sustaining, because the existing organizational structures and processes do not support the innovation. However, for compatible innovation, an independent entity may not be necessary even if the innovation is disruptive. With enough resolve, an incumbent’s upper management can overwrite the priorities for decision-making and resource allocation that would otherwise lead to the rejection of a disruptive innovation. Two prominent examples:

Intel Celeron — Andrew Grove, then-chief executive of Intel, led Intel to develop the Celeron chip after a discussion with Clayton Christensen about the hazards of disruptive innovation. The Celeron chip was a lower-performance, lower-priced alternative to Intel’s flagship Pentium chip, and therefore a disruption to Intel’s own core business. But even though the Celeron was disruptive, Intel was perfectly capable of developing it and bringing it to market without an independent entity, because the Celeron was a compatible innovation and did not require material alterations to Intel’s existing organizational structures and processes.

Battery-powered outdoor tools — Powered outdoor tools like chainsaws, hedge trimmers, brush cutters, and leaf blowers have traditionally been built with combustion engines. But over the past decade, a new category of battery-powered outdoor tools with electric motors has seen growing popularity. Battery-powered outdoor tools have numerous advantages over combustion-engine-powered outdoor tools. They cost less, weigh less, are less noisy, are easier to operate, and require less maintenance. But battery-powered outdoor tools have downsides, too: They cannot (yet) compete with their combustion engine counterparts in terms of performance and runtime in the field, two key purchasing criteria for many customers. Battery-powered outdoor tools are a disruptive innovation because they are entering the outdoor tool market from the low end. They are also a compatible innovation: The majority of incumbents today offers a portfolio of battery-powered outdoor tools that was developed and commercialized without independent entity. The incumbents were able to do this because the organizational structures and processes required were similar to those they already had in place.

And the list goes on. In fact, three key examples for disruptive innovation from Clayton Christensen’s book The Innovator’s Dilemma — 5.25-inch hard-disk drives, hydraulic excavators, and mini steel mills — are compatible innovation. As the theory of compatible versus incompatible innovation predicts, all of them were eventually adopted by incumbents, without independent entity:

5.25-inch hard-disk drives — 5.25-inch hard-disk drives were disruptive to 8-inch hard-disk drives. They could be used in personal computers thanks to their smaller size and a form factor that fit in the slot for a 5.25-inch floppy drive. But their lower capacity and higher cost per megabyte made them unattractive for minicomputers, the mainstream market for 8-inch hard-disk drives back then. Seagate, a startup at the time, released the first 5.25-inch hard-disk drive in 1980. Other players followed quickly when 5.25-inch hard-disk drives took off due to the rapid adoption of personal computers, including incumbent Control Data Corporation. A leader in 8-inch hard-disk drives, Control Data Corporation built the first high-performance 5.25-inch hard-disk drive, the Wren. Because 5.25-inch hard-disk drives were a compatible innovation, Control Data Corporation was able to develop and market the Wren without independent entity.

Hydraulic excavators — Hydraulic excavators were disruptive to cable-operated excavators. They had numerous operational advantages over cable-operated excavators, such as greater sturdiness, a better motion range, and an ability to work at higher speeds without shocks or damage. On the flip side, they were initially less powerful and constrained to niches like drainage or landscaping, whereas cable-operated excavators continued to dominate larger projects like building construction. When hydraulic excavators became more powerful and expanded into the mainstream market, several of the incumbent manufacturers of cable-operated excavators adopted the technology and ultimately discontinued cable-operated excavators. For example, Hitachi sold cable-operated excavators since 1949, entered the market for hydraulic excavators in 1965, and is today a leading manufacturer in this market. What made it possible for them to do this without independent entity was the fact that hydraulic excavators were a compatible innovation.

Mini steel mills — Mini steel mills were disruptive to integrated steel mills. They are used to produce lower-quality steel by melting down scrap metal. The process is faster, takes fewer man-hours per ton of steel, and is less capital-intensive compared to operating integrated steel mills. But because the quality of the steel is lower, the range of products that can be produced is more limited as well. Some of the large steel makers, like United States Steel Corporation, never adopted mini steel mills. But others did. For example, the world’s largest steel producer, ArcelorMittal, today is operating both integrated steel mills and mini steel mills. In addition, Armco, which since changed its name to AK Steel, operated two mini steel mills in the past. While one never reached profitability, the other became one of Armco’s most profitable steel mills at the time. Both examples show that incumbent steel producers were able to take advantage of mini steel mills without independent entity because mini steel mills were a compatible innovation.

Clayton Christensen’s theory of sustaining versus disruptive innovation is seminal, as it predicts with remarkable accuracy which innovations corporations recognize as threats and which they choose to ignore, often until it is too late to catch up. The theory of compatible and incompatible innovation proposed in this article complements his work in that it helps explain which innovations corporations can successfully adopt, specifically without setting up an independent entity, and which they will struggle to respond to, regardless of whether these innovations are sustaining or disruptive. Corporate innovators should heed both theories because they will help them understand which innovations to focus on, and how to go about them.

Premainstream

Corporate innovation, hands-on: Best practices to heed, pitfalls to avoid, and great case studies to take as examples.

Christian Vogt

Written by

High-impact, fast-paced, risk-controlled corporate innovation — inside-out and outside-in.

Premainstream

Corporate innovation, hands-on: Best practices to heed, pitfalls to avoid, and great case studies to take as examples.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade