The Hardware Startup

Chapter 1. The Hardware Startup Landscape

If you’re reading this book, it’s likely because you’ve decided to start, or are thinking about starting, a hardware company. Congratulations! Launching a hardware startup is an exciting and challenging undertaking. There’s a saying: “Hardware is hard.” You have to navigate the complexities of prototyping and manufacturing, the daunting optimization problems of pricing and logistics, and the challenges of branding and marketing. And you’ll be doing it all on a pretty tight budget.

But today—right now!—is probably the best time in history to be starting your company. Technological advances, economic experiments, and societal connections have facilitated the growth of an ecosystem that enables founders to launch hardware companies with fewer obstacles than ever before.

Before we get into the specifics of getting your business off the ground, let’s set the stage by discussing some important influences that have brought the ecosystem to where it is today.

Early Makers

Today’s hardware entrepreneurs stand on the shoulders of early makers. The maker movement has had a profound influence on the hardware-startup ecosystem. Defined by three characteristics—curiosity, creativity, and community—it emphasizes project-based learning, learning by doing, and sharing knowledge with others. Experimentation is important. Having fun is a priority.

While people have always had a desire to make things and work with their hands, the rise of a distinct hobbyist do-it-yourself (DIY) culture focused on technology began in the 1960s.

The Whole Earth Catalog

Stewart Brand’s Whole Earth Catalog, which first appeared in 1968, was one of the foundational resources of what became the maker movement. More than just a catalog, it was a manual for people who wanted to live a creative, DIY lifestyle, and a cornerstone of 1960s counterculture. Tools, machines, books, farming products—all of these could be found in the catalog, along with vendor names and prices. Customers could buy directly from manufacturers.

The catalog featured how-to guides on everything from welding to breeding worms. The emphasis was on personal skill development, independent education, and what’s now called life hacking. John Markoff, technology writer for the New York Times, referred to it as “the internet before the internet” and “a web in newsprint.” It captured the imaginations of a generation of counterculturalists, many of whom went on to careers in technology.

The catalog ceased regular publication in 1974 and was published intermittently until 1998. The back cover of the last regular edition had a farewell message: “Stay hungry. Stay foolish.” This famous phrase is often attributed to Apple founder Steve Jobs, who called the Whole Earth Catalog “sort of like Google in paperback form” in his famous 2005 Stanford commencement address.

Communities Around New Technology

In the 1970s, computers captivated the imaginations of many early Silicon Valley technologists, including Steve Jobs. Communities sprang up around the new technology. One example was the Homebrew Computer Club, a group of Silicon Valley engineers who were passionate about computers (particularly early kit computers). Members included Steve Wozniak and Steve Jobs. The club, which met from 1975 to 1986, was instrumental in the development of the personal computer. Wozniak gave away schematics of the Apple to members and demoed changes to the Apple II every two weeks.

These early adopters took the DIY ethos of the Whole Earth Catalog and extended it to DIWO (do it with others). At first, software was the primary beneficiary of this collaborative spirit. The free and open source software movements, which advocated the release of software whose source code was public and modifiable by anyone, began in the mid-1980s and steadily gained in popularity.

By the mid-1990s, the trend moved from bits to atoms, and an open source hardware movement began to grow (see “Open Source Hardware” for more information and examples). Open source hardware is “hardware whose design is made publicly available so that anyone can study, modify, distribute, make, and sell the design or hardware based on that design.”

MIT Center for Bits and Atoms

By the late 1990s, maker culture and prototyping technologies were becoming more formalized in academic institutions. Often called the “intellectual godfather of the maker movement,” Neil Gershenfeld founded the MIT Center for Bits and Atoms (CBA) in 2001. The CBA focuses on creating cross-disciplinary fabrication facilities that offer shared tools, with the intent to “break down boundaries between the digital and physical worlds.”

These fab labs are scattered around the world, but they share core capabilities that allow people and projects to move freely among them. Projects range from technological empowerment to local problem-solving to grassroots research. Several prominent companies with a distinct maker ethos and strong ties to the community have emerged from work done at the CBA, including Formlabs, Otherlab, Instructables, and ThingMagic.

Make Magazine

As community-driven innovation and small-scale fabrication experiments were taking root in academia, the maker movement was steadily gaining popularity among hobbyists. DIY pursuits were steadily becoming more mainstream. Dale Dougherty, cofounder of O’Reilly Media and developer/publisher of the Global Network Navigator, noticed the increasing interest in physical DIY projects among his peers in the tech community.

Dale had previously created the Hacks series of books for O’Reilly. The series helped users explore and experiment with the software they used, empowering them to create shortcuts and useful tools. In 2005, Dale created Make magazine, based on a related, simple premise: “If you can mod software, you can mod the real world.” He and O’Reilly Media cofounder Tim O’Reilly had spent years enabling people to learn the skills necessary to write software. Make was conceived to help them learn the skills needed to make things in the physical world.

In addition to teaching practical skills, Make emphasizes creativity. In 2006, the Make team put on the first Maker Faire, with the goal of bringing the maker community together in person to showcase and celebrate the DIY spirit. Dale remembers:

I noticed that a lot of really interesting work was happening in private. We see objects every day, but there’s nobody talking about how they were built. I wanted to create a place for people to have conversations about that in public, in a way that was enjoyable and fun.

The 2006 event had 200 exhibits and drew 20,000 attendees. By 2013, there were 900 exhibits and 120,000+ attendees. The original flagship Maker Faires were held annually in San Mateo, New York City, and Detroit, and their numbers have exploded, with new Faires launching every year in Europe, Asia, and around the US. Community-run Mini Maker Faires have popped up around the world as well.

As maker culture has become increasingly popular, thousands of people have been inspired to create unique projects that solve personal pain points or provide entertainment. Community hackerspace founders have taken the fab lab model and used it as inspiration for shared neighborhood workspaces. In addition, the rise of the Internet enabled communities to form unencumbered by geographical distance (see “Online Community” for more information and examples). Technology enthusiasts from all of the world can connect with one another and share.

Technology Enables Scale

Over the past five years, we’ve begun to witness the emergence of maker pros: entrepreneurs who started out as hobbyists and now want to turn their creations into full-fledged companies.

The difference between a project and a product is the difference between making one and making many. To turn a project into a company, the product has to be scalable. “Making many” has traditionally been a problem of cost and accessibility; it’s historically been both expensive and difficult to manufacture. Growing a company further requires keeping costs low enough to profit, setting up distribution channels, and managing fulfillment.

Over the past few years, several trends have combined to create an environment that’s mitigated those problems. This has resulted in the growth of a hardware startup ecosystem.

Rapid Prototyping

Advances in rapid prototyping technologies have fundamentally changed the process of taking an idea from paper to the physical world. Hobbyist and prosumer-level 3D printers, computer numerical control (CNC) routers, and laser cutters have altered the landscape of personal fabrication, enabling quick and affordable iteration.

While 3D printing has been around since the 1980s, the cost of a machine has dropped dramatically. Materials such as metals and ceramics enable higher-fidelity models. Cloud-based fabbing services, such as Ponoko and Shapeways, can produce a single prototype and ship it to you within a week—no need to own the printer yourself!

Inexpensive boards (such as Arduino, Raspberry Pi, and BeagleBone) make electronics prototyping accessible to everyone. As interest in the Internet of Things has grown, products such as Spark Core and Electric Imp (startups themselves) have hit the market to make connected-device prototyping fast and easy.

Simultaneously, computer-aided design (CAD) software has become more sophisticated, more affordable, and easier to use.

Inexpensive Components

Just as the cost of major prototyping technologies has come down, component prices for sensors, batteries, and LEDs are also much lower. Several early maker businesses (MakerBot, Adafruit, SparkFun) are excellent resources for prototyping supplies and technologies.

Ubiquitous smart devices have also had a dramatic impact on the hardware ecosystem. Global smartphone penetration is at 22 percent; within the United States, it’s at 56.5 percent and growing (and reaches as high as 70 percent in some countries). While this phenomenon is partially responsible for driving down the cost of component parts, the smartphone itself has also had a dramatic impact on hardware devices. It’s an increasingly common interface through which humans can interact with connected devices and wearables.

Small-Batch Manufacturing

As machine costs have dropped, small-batch manufacturing has become increasingly feasible. The minimum number of units needed to secure a contract manufacturer used to be in the tens of thousands, but today’s factories are increasingly willing to do small-batch runs (sometimes in the hundreds of units).

Small batches are one way that a fledgling hardware company can run lean. Even if software-style constant incremental iteration is still impossible, the amount of money lost to a bad run is considerably reduced with a small batch. Increasing awareness of Shenzhen’s growing manufacturing ecosystem and increased ease of sourcing through sites such as Alibaba and Taobao have also opened up China as a viable option for smaller startups.

Open Source Hardware

Open source hardware platforms are continuing to gain popularity, allowing entrepreneurs to build on top of them. Arduino, for example, eliminates the need to build a proprietary board during the early development phase.

As of 2011, there were more than 300 open source hardware projects, and the number continues to increase. Engaged communities of contributors help accelerate innovation, and their accessibility and willingness to share knowledge draw in new makers. This democratizes innovation.

Open source can also be a business in itself; MakerBot and Arduino are thriving companies in their own right. By 2010, each already had over $1 million in revenue, and MakerBot was acquired by Stratasys in 2013 for $604 million.

The Open Source Hardware Association (OSHWA) is the present-day voice of the open source hardware community. It works to advance the goals of collaborative learning and promoting the use of open source hardware.

Online Community

In addition to generating awareness of the hardware space and helping people learn more effectively, community knowledge-sharing helps spread best practices and innovative ideas. Web-based communities such as Instructables and Thingiverse are geography-agnostic; they enable people around the world to share projects online and learn from others. Sometimes communities come together to contribute funding to a particular project. Crowdfunding platforms help founders leverage community support and bring products to market.

While online communities can provide support and access to information, geographically concentrated local communities can help members overcome design and prototyping challenges by making access to expensive machinery much more feasible.

Hackerspaces (often called makerspaces if the primary focus is physical device hacking) are local hubs for hobbyists, crafters, and maker pros alike to come together and build things. Beyond creating a welcoming physical space and facilitating collaborative serendipity, they often include shared tools and machines (similar to the MIT Fab Lab model discussed in “MIT Center for Bits and Atoms”).

Some hackerspaces are simple community garages. Others, such as TechShop, involve paid memberships and offer courses for skill development. These spaces harness the power of sharing, creating an “access-not-ownership” model that makes even expensive professional-grade equipment relatively accessible. Over the past five years, makerspaces and hackerspaces have spread across the world. The Hackerspace wiki, which tracks spaces globally, lists over 1,600 spaces. Many are primarily devoted to software, but a steadily increasing number focus on hardware.

The Supplemental Ecosystem

Getting a device made is only the beginning of a successful hardware endeavor. To turn a project into a company still requires navigating fundraising, inventory management, distribution, customer service, and more.

New businesses are being launched specifically to help hardware startups navigate these challenges. Accelerators (Chapter 7) that traditionally offered funding, mentorship, and assistance to software companies are expanding into hardware. Hardware-specific programs are popping up, providing the specialized assistance necessary for startups to efficiently produce physical goods; some focus specifically on helping startups navigate manufacturing overseas. Fundraising platforms such as Kickstarter and Indiegogo can help entrepreneurs validate markets, raise money, and grow engaged communities (for more details, see “The Crowdfunding Ecosystem”).

Once the product has been made, fulfillment-as-a-service shops enable entrepreneurs to offload some of the logistical challenges of warehousing, packing, and shipping. Distribution channels, such as Grand St. (recently acquired by Etsy), Tindie, and ShopLocket provide a means to easily reach consumers without needing to go through big-box retailers.

The “Lean Startup” and Efficient Entrepreneurship

The templatization of best practices for software startups has had a profound impact on entrepreneurship. The Lean Startup movement, introduced by Eric Ries in 2011, is a series of best practices designed to make starting a company a more feasible, less risky proposition for aspiring founders. Lean startups identify clear customer needs and incorporate customer feedback into the product design from day one, iterating rapidly to produce a truly useful product. They are strongly grounded in data-driven assessments of their offerings, using techniques such as A/B testing and closely monitoring actionable metrics (metrics that tie user data to outcomes). While running a lean hardware startup is fairly challenging, the popularity of the movement has inspired thousands of individuals to think seriously about turning their projects into companies.

One of the core principles of the Web 2.0 movement was that everyone is capable of “being a creator.” Online, that spirit has been reflected in the rise of blogging, photo taking and sharing, pinning, tweeting, and creating web content. In the physical world, “being a creator” means making physical goods.

Dale Dougherty compares this progression from makers to entrepreneurs to a similar phenomenon that happened in the early days of the Web:

Early on, most people were creating websites because they could. At some point, people said, hey, there’s a way to make money from this—I’m not building websites; I’m building a way to make money.

The maker movement has increased the pervasiveness of the DIY spirit, facilitated easier access to information, and generated a supportive community that helps today’s founders get companies off the ground. It has helped millions of people realize that they, too, can hack the physical world. People start off small: they learn how to make one of something. But once they’ve made one, making many—and starting a company—no longer seems like an impossible task.

The Hardware Companies of Today

Most hardware startups make products that fall into one of four subcategories: connected devices, personal sensor devices, robotics, and designed products. Admittedly, some hardware falls into multiple categories; your phone, for example, is a personal sensor device (it has an accelerometer and gyroscope that can be used to measure the activity of the person carrying it). When you open an app for your smart watch or fitness tracker, it becomes part of a connected device. And if paired with a product like Romo, it can even become a robot.

Some products are difficult to classify, but for the purposes of this book, we’ve found that these categories make the most sense for discussing the challenges of bringing certain types of products to market.

Connected Devices

The term connected device broadly refers to a device that has a cellular, WiFi, or other digital connection but is not a cell phone or personal computer. Some of these devices (ereaders, tablets) are designed to be used by people. However, the term is increasingly used to refer to devices that are connected to, and communicate with, other machines (machine-to-machine, or M2M). A growing number of connected-device hardware startups fall into this category. They are the startups that are building the Internet of Things.

The term Internet of Things was originally coined by Kevin Ashton, cofounder of the Auto-ID Center at MIT. Ashton recognized that most of the data on the Internet was gathered or created by humans:

Conventional diagrams of the Internet…leave out the most numerous and important routers of all—people. The problem is, people have limited time, attention, and accuracy—all of which means they are not very good at capturing data about things in the real world. And that’s a big deal.

The broad vision for the Internet of Things is a world in which objects connect to the Internet and transmit state information without human involvement. It’s quickly becoming a reality. Cisco Systems, a manufacturer of networking equipment, states that as of 2010, there were 12.5 billion devices connected to the Internet: “more things than people.” By 2020, projections from Cisco and Morgan Stanley estimate that 50–75 billion devices will be connected.

Internet of Things objects use networked sensors to generate data, which is then analyzed by other machines. The objects themselves can in turn be modified or controlled remotely. While many such devices have a consumer focus, the promise of the Internet of Things extends to Big Industry as well. Connected systems form the backbone of the Industrial Internet, in which identifiers, sensors, and actuators work together to form complex autonomous systems in industries ranging from manufacturing to health care to power generation. The specific benefits vary—some industries are interested in cost reduction, while others care about improved safety—but the promise of the Internet of Things is one of better outcomes, improved by increased productivity and efficiency.

Making homes and cars “smarter” is a popular consumer vision. Startups such as Nest (recently acquired by Google), August, and Automatic are working on connected smoke alarms, thermostats, door locks, and vehicles. Others, such as SmartThings, are producing beautiful dashboards that act as a hub for monitoring connected home devices.

Given the vast market potential, many large companies have entered the IoT space. Belkin’s WeMo plugs into electrical outlets and enables a smartphone to control the outlet (and the device plugged into it). In a new partnership with appliance maker Jarden, the WeMo Smart can turn Jarden’s Crock-Pot and Mr. Coffee lines (among others) into connected and controllable devices. Lowe’s produces the Iris Smart Home Management System, which offers sensors for security, temperature control, power management, and more, which all transmit information back to the owner’s smartphone. In some cases, the human user need not be consulted; garden-soil monitoring sensors can detect dryness and automatically trigger the watering system.

Asset tracking is another popular application. Mount Sinai Hospital in New York has begun tracking assets (e.g., hospital beds, wheelchairs, and pain pumps) with radio-frequency identification (RFID) tags. Large farms are also getting in the game, marking cows with RFID tags to track when the animals feed and how much milk they produce.

This sector of the hardware ecosystem has benefited extensively from low-cost sensors and ubiquitous smartphones. Large-scale data-processing techniques have also had a profound impact; the ability to turn vast quantities of information into meaningful insights is increasingly important as more devices become connected. It’s a great space to be building a business in. We’ll be focusing on the pitfalls unique to hardware startups that make connected devices, such as security, standards, and power management. We’ll also cover producing a seamlessly integrated software experience, ease of use, and competing with very large incumbents.

Personal Sensor Devices (Wearables)

The line between personal sensor devices and more general “connected devices” is somewhat blurry. The connected devices mentioned in the previous section are defined by their ability to autonomously communicate with other devices, and many wearables do exactly that.

For our purposes, “personal sensor devices” will refer to products that gather data related specifically to a human subject, then process and display it in a way that makes it easily understandable to human end users. This typically takes the form of a device (frequently worn by the subject) and a mobile app or dashboard that presents logs or visualizes trends in the data. A smart watch, for example, may track wearer step counts and sync to an app on the user’s mobile phone.

The market for personal sensors and wearable devices grew organically out of the Quantified Self movement, which focuses on tracking personal data. As far back as the 1970s, people were experimenting with wearable sensors, but the movement began to gain mainstream attention around 2007. Gary Wolf and Kevin Kelly began featuring it in Wired magazine, and in 2010, Wolf spoke about it at TED. Since Quantified Self was a movement largely led by technologically savvy early adopters, it’s not surprising that much of its early activity came from startups.

Health and wellness are the primary focus of most of today’s wearable sensor devices. Activity monitors, which are designed to help people become more aware of their fitness practices, are the most common. Other wellness-device startups are attempting to tackle sleep tracking, weight monitoring, dental hygiene, and brainwave measurement.

On the diagnostics side, startups are pursuing blood glucose monitoring, smart thermometers, mats that alert diabetics to foot ulcers, and “smart pills” that monitor compliance. These applications portend a future in which the medical industry is increasingly reliant upon sensor technology. These devices often require some degree of US Food and Drug Administration (FDA) approval, which we will touch on briefly.

Outside of health and wellness, a broader wearables market has emerged with applications in fashion, gaming, augmented reality, lifelogging, and more.

The increased prevalence of smartphones, better battery life, low-energy Bluetooth connectivity, and the dropping cost of sensor production have made this an attractive space to build a company in. Widespread public adoption, particularly in wellness and fitness, has made the consumer market more attractive. Social networking and interconnectedness have driven user adoption, as friendly competition and data sharing help people set goals and stay motivated.

As in any growing space, big companies are paying more attention to the market in personal sensor devices. Nike has long offered run-tracking technology in the form of the Nike+ (a sensor that connects to a runner’s shoe). In 2012 it expanded into the FuelBand, a bracelet that constantly monitors daily activity levels, though in early 2014 it announced that it was no longer producing the product. Reebok and MC10 have partnered to develop an impact-sensing skullcap for athletes in contact sports. UnderArmour’s Armour39 is a chest strap and module (with an optional watch) that measure heart rate, calories burned, and general workout intensity.

These devices are a departure from the core business of these companies, but the combined temptation of a large addressable market (see the entry for “Addressable Market” under “Telling a Story”) and a desire to be seen as cutting-edge has led them to push the envelope. There’s also the data; while the users benefit from it, the device manufacturer is learning as well…about the habits of its customers.

User experience and user interface design are particularly important considerations when building a personal sensor startup, and so is privacy. Data control is an issue that founders must keep in mind when building a business. We’ll touch on these factors throughout the book.

Robotics

The third subset of hardware startups is the robots. These automated machines are designed with an eye toward improving the lives of humans. Some, such as home cleaning robots or robotic “pets,” just make everyday consumer life a bit more convenient or fun. Others are used for important tasks in industry, such as improving the efficiency of the assembly line or doing hazardous work such as defusing bombs. The wide-ranging applications and demand for robots spans the consumer, military, and commercial markets, making it an extremely lucrative space.

Autonomous robots are a relatively new technology, appearing in the second half of the twentieth century. The first, the Unimate, worked on a General Motors assembly line. Its job was to transport molten die castings into cooling liquid, and weld them to automobile frames. The Unimate was a large stationary box with a movable arm, but it did important work that was too dangerous for humans. Since the Unimate, robots have changed industry in three primary ways:

• Their accuracy, consistency, and precision have improved product quality.

• Their ability to do work that humans shouldn’t, or can’t, has made manufacturing safer.

• Their cost relative to value—particularly when factoring in increased productivity—has fundamentally altered the bottom line of many companies.

Robots now roll, balance, swim, sail, climb, and even fly (drones). Sensor technologies have enabled tactile, auditory, and visual input processing. Interaction experts are working to perfect the user experience of interfacing with robots. Improved actuator technology has reduced costs and size, and advances in computation have expanded the types of activities that robots can perform (and their degree of autonomy). It’s a rapidly growing industry that continues to push the limits of what’s possible in many other fields.

While large factory-floor robots are still primarily developed by big companies, there are a number of startups working in the industrial space. One example is the Baxter robot by Rethink Robotics, which is designed to work with humans on assembly lines without a safety cage. Fetch Robotics, founded in 2013 as Unbounded Robotics, is building a human-scale mobile manipulator robot at a $35,000–$50,000 price point, a fraction of the cost of similarly featured products.

Robots are gaining popularity off the factory floor too. They are increasingly being used in agriculture and farming; agricultural robots from Spanish startup Agrobot thin lettuce and pick strawberries. There are underwater robot (ROV, or remotely operated vehicle) startups such as OpenROV, and flying robot (unmanned aerial vehicle, or UAV) startups such as 3D Robotics and DroneDeploy.

In the consumer market, startups are tackling use cases ranging from telepresence to children’s toys. Health care and home assistance are becoming popular sectors.

Like the hardware verticals mentioned previously, robotics has its own unique business and manufacturing challenges. Robotics products are often costly to manufacture, and finding the right partner can be particularly challenging. This can pose a challenge for a startup when it comes to identifying a go-to-market strategy, because it can be difficult to get the costs down to the point where the price is appealing to consumers.

Robotics startups have been particularly popular acquisition targets. In 2013, Google alone acquired eight robotics startups.

Designed Products

Our fourth category, designed products, is an admittedly broad catchall. These companies make purely physical devices; the startup is generally building hardware only, with no software. Some are simply made things; we’re talking everything from 3D-printed custom dolls for kids to the kitchen gadgets sold in Bed Bath & Beyond.

A pioneering startup called Quirky changed the face of bringing a designed product to market by incorporating community. In Quirky’s process, inventors submit ideas, and the community curates the ideas, provides feedback, and eventually votes on the market potential of ideas that gain a following. If the process goes favorably, the community continues to assist with research, design, and branding, and ultimately helps the inventor arrive at a final engineered product.

Following manufacturing (also handled by Quirky), the product moves into sales channels. This is a combination of direct and social sales, plus partnerships with retailers. The profit from the product is shared among the community as well, according to the impact each person made in getting it from thought to market. Quirky keeps 60 percent.

We chose to use a four-category framework in this book, because we’ll occasionally point out business or engineering considerations that are specific to one particular subset. But generally speaking, the number of hardware startups across all categories has increased as the old barriers of large capital requirements and long lead times have largely fallen away.

So, to summarize: hardware is getting less hard. Increasingly available prototyping tools, such as 3D printers and laser cutters (which are themselves frequently the offerings of hardware startups), are enabling entrepreneurs to apply “lean” startup principles (see “The “Lean Startup” and Efficient Entrepreneurship”) to hardware companies. A growing ecosystem of support companies is helping to reduce the traditional complexities of marketing, inventory management, and supply chain logistics. We’ll cover all of those factors in depth in this book, as we work through a roadmap for The Hardware Startup.