Sunday, June 17, 2012

Is hardware the new software?

When I observe what’s happening to consumer electronics I get a feeling of déjà vu.  Is consumer hardware going through the same transformation that consumer software did in the last decade? Change has been simmering in the hardware space for years and now I wonder if we are about to witness a convergence that will cause transformational shifts and disruptions to hardware like those brought about in software with the advent of “the cloud”?

Ten years ago if you had a good idea and raised a million dollars in seed funding to start an Internet software company, you’d spend 20-30 percent of your capital, sometimes more, on basic infrastructure. You had to rent your own hosting cage. You had to lease and rack-and-stack your own servers, switches and firewalls. And you would likely have to license and install costly software. Finally you had to hire someone to manage it all.

There was no app store and no Facebook “likes”. You had to manufacture physical discs and boxes and shrink-wrap the software. You had to do "distribution deals" with retailers. And you needed a co-marketing budget in order to buy ads and shelf space.

It was not uncommon for a company to raise a full $3-5 million before even launching their product or service. Efficiency of capital in the software space was low ten years ago because so much of it had to be spent on undifferentiated, commodity infrastructure and costly awareness building.

The consequence of having to raise a lot of capital in order to build software was that barriers to entry were high and innovation stifled. Ten years ago it was almost impossible for a 22-year old college grad to start a software company, regardless how smart they were or how good their idea was, because in order to get going you needed to enter into business relationships and secure multi-year server and software leasing contracts. You had to hire a management team to demonstrate to both investors and vendors that you had what it took to be successful.

All that has changed. What enabled this change was the emergence of a utility model for computing and storage, an operating system model that encourages a long-tail of innovation through apps, plus a ubiquitous network that ties it all together with customers.

Today all you need to build a web service is a credit card and an Amazon Web Services (AWS) account. Or with one or two engineers you can build a mobile app for iOS, Android or Windows Phone and get instant distribution through an app store. You can hardly get your hands on $100k in seed funding without having a working prototype of the thing you’re building.  It's not easy to stand out in a sea of online services and apps, but the cost to take a shot at it has dropped like a rock in a pond. Anyone can run the experiment. Access to capital is no longer the primary gating factor when trying out a software and services idea.

The way we develop and distribute software has radically changed as a primary function of five pervasive and loosely coupled developments: open source software, open services and APIs, “the cloud” aka hosted computing and storage, primarily Amazon Web Services, mobile platforms with app stores and finally access to the social graph for awareness building and distribution.

While a small amount of effort is still required to set up and configure the servers and software to run a new service, the vast amount of work now goes into the development – coding – of the core idea, failing early and iterating or moving on if the experiment fails.

The blossoming and explosion of web and mobile apps in the last ten years are a direct consequence of major barriers to innovation disappearing. The reduced need for capital has had a profound democratizing effect. DropBox, Angry Birds, Instagram, even Facebook are a few high profile examples where a few people have created hugely successful applications and services with minimal amounts of start-up capital. These examples are only a few exceptionally bright beacons in a vast, seething ocean of creativity and innovation, all fueled by ubiquitous access to “the cloud” and to simple distribution through “app stores”.

Are we going through a similar transformation in hardware?

A couple of weeks ago Wired ran a story called In the KickstarterFuture, Hardware is the New Software.  They featured two guys who used Kickstarter to raise only $32,000 to manufacture their first batch of an air guitar pick that plugs into the iPhone. They had raised a little bit of money from angel investors first in order to design the product, but altogether were able to build and ship their first product for just over $100,000. This would have been unheard of just a few years ago.

On May 18th a small start-up called Allerta became the talk of the tech community by setting a new record on the crowd-funding site KickStarter. They raised an astonishing $10.2 million from individuals who pre-ordered, and hence funded manufacturing of the $150 Pebble smart-watch. Pebble is a cool looking Android based smart watch that wirelessly connects to your smartphone and shows alerts from your phone, including caller ID, email messages and Facebook notifications. It even has APIs that lets developers write their own apps that run on your watch.

 The Pebble Smart-Watch pitch

Yet the most interesting part of the Pebble story, a part that most people seem to have missed, was not that they raised more than $10 million. The astonishing thing was they were only trying to raise $250,000!

With only a quarter of a million dollars a small company with a few young guys – five if I’m not mistaken – were planning to build and ship their first fully functioning, nicely designed, Android based smart-watches. In fact, when they posted their pitch to Kickstarter they already had a working prototype that they had built based on learning from a prior, similar product called InPulse that they had built for BlackBerry phones.

How is it that a few talented guys – used gender neutrally of course – with very little capital can now build and sell functioning hardware products? Is there a “cloud equivalent” that is changing the rules of the game in hardware too?

While the cloud is an apt metaphor for the “bits of software”, it is not the right model to describe the “atoms of hardware”. So perhaps we should be calling what’s happening to hardware an “earth infrastructure” as opposed to the software “cloud infrastructure”. Something big is going on with this earth infrastructure and here are the drivers:

3D printing and modeling tools – a remarkable revolution is underway in 3D printing. Physical objects can literally be designed on a laptop and printed as full physical models that we can touch, feel, and see in the real world. 3D printing technology is on a roll and it is having an effect on both perception and reality of how “easy” it has become to produce a physical object. “Just hit print” and at rapidly falling cost you can pick up a model and see if it resembles what you had in your mind. This is having a big effect on the speed and cost of iterating on designs and form-factors.

Maker movements and open source – The Maker movement today looks a lot like the early open source software movements. The Maker movement still has a decidedly “hobbyist” flavor, yet growing rapidly in breadth and depth. Hardware incubators such as Lemon Labs based in SF is bringing the incubator model that's blossomed in the software space to hardware. And the open source model is emerging as well. Adruino and Bug Labs are both open source electronics platforms. Thingiverse is emerging as an interesting repository of digital designs. While a full-on GitHub equivalent in the hardware space has not emerged yet, it’s likely that this will happen.

Android – while the lower layers of the operating system has been commoditized for a while, Android is rapidly making the whole OS stack a commodity. Equally important is the fact that the success of Android as a mobile phone OS has made it much, much easier than it was before to get set up and running on virtually any chip set. All chip vendors have an Android implementation and apps will (more or less) run on all of them. Just like people used to put a lot of effort into the hosting infrastructure pre-cloud, only a few companies had access to an OS. Others would license it. Owning the OS was a significant barrier to entry. Not so much any more. With Pebble a couple of guys were able to take Android and get it up and running on their smart watch. The equivalent would have been very expensive or even impossible even two or three years ago. Today someone wanting to build a “smart gadget” can focus most of their effort on the user experience and apps, as opposed to investing heavily in developing or significantly customizing the OS.

Chipsets – from sensors to CPUs with micro controllers in the middle, it is now possible to get powerful, fully functional computing devices up and running in record time. Reference designs that run Android are readily available by all the chip manufacturers leading to a significant reduction in cost and complexity. Much like what happened with internet software development, a much larger part of the development effort can now be focused on creating end-user value, rather than developing a low level software stack and designing chip sets.

Shenzhen – a powerful hardware manufacturing eco-system has emerged in Shenzhen, China. And as pointed out in the Wired article, the Chinese manufacturers are now increasingly willing to take on small start-ups. The competitiveness of the eco-system leads to constant improvements in quality and pricing. And this also leads to efficiencies where it is no longer necessary to start with large orders. It has become possible to start a production run with only a few thousand devices, again significantly reducing the need for start-up capital to get a consumer electronics product off the ground.

Two other trends supporting hardware innovation are access to capital through crowd sourcing services such as Kickstarter and the pervasiveness of ecommerce options, reducing dependencies on traditional retailers and distribution channels. Even if a product still needs to make it into traditional retail channels in order to make it big, it is possible to get going, get market validation and get your first customers by selling direct, asking your first customers to help promote your products on Facebook and Twitter, circumventing traditional dependencies on big marketing budgets and distribution agreements at the beginning of a new product’s life.

The “earth” – as opposed to “cloud”, remember – manufacturing and distribution infrastructure necessary to support design and manufacturing of atoms – aka hardware – is becoming very efficient resulting in the barriers to entry in hardware coming down fast.

What is happening in consumer electronics looks a lot like what has happened in software in the last ten years. A small team can now make hyper efficient use of capital and voila, out comes a cool little physical object, manufactured with earth infrastructure, running a state of the art operating system with apps, and obviously connected to a hyper efficient cloud infrastructure.

If I were a venture capitalist today I’d be investing in next generation earth infrastructure!

No comments: