“Makers are enthusiasts who hack and modify the world around them in interesting and whimsical ways. Tools and services that used to be inaccessible to all but large manufacturers are now available to everyone. Foreign factories that were impenetrable before are now an email away. Design software costing thousands of dollars per seat is freely available (or very cheap). Hackers are mixing all of these elements together and re-imagining entire industries from the ground up.”
– Vinod Khosla, The Unhyped New Areas in Internet and Mobile, Techcrunch, February 19, 2012.
Vinod Khosla is right. The time is ripe for startups working on products combining hardware, software, and the internet.
Those who know me know that I have been tooting the HW+SW+Internet horn since early 2010. I hinted at it in my post Internet Startup Bubble and The Supposed Super Seed Crash back in July of 2010. But now, more than ever, the evidence is here that hardware is ready to make a comeback in startups. In fact, it already is.
But it’s been hard talking about this for the last 2 years. I’ve met nothing but resistance from the investor community. Practically every VC I’ve talked to has told me that hardware is awful and that software and internet is much better as an investment. And for a long time, it’s been true.
If you think about it, the last VC who was OK with investing in hardware startups entered into the VC business in 1994. After that, the internet came into being and from that point forward, every VC who didn’t jump on the internet bandwagon missed one of the biggest opportunities to make a ton of money. (Of course, those that didn’t jump off the bandwagon at the right time lost their shirts and more). Still, memories about the downsides of bubbles are short; those same VCs were heroes for the money they made and they continued to invest in the internet, and training hordes of emerging VCs along the way. This has continued for 17 years now. Think about it; 17 years of VCs whose thinking has been molded by the success of internet startups and investing in them. And also the advantages of not having to build physical product to get there.
I think the world has changed to the point where the previous advantages of internet/software startups has been declining, and the advantages of hardware startups are ascending.
For a long time, internet/software startups had a distinct advantage over hardware startups. You didn’t have to use up money in paying for inventory of product. Digital products cost so much less and upon copying the bits digitally, the cost of the product declined over time as you sold more. Plus, the internet created distribution mechanisms that were hard to compete with; customers could be reached with extremely low cost and sold to with great ease. But that has changed:
1. Competition in internet/software startups is way too fierce. You start something and competitors pop-up all over the place. With all the easy ways to create software, it is extremely easy to build something that somebody else has built and do it fast.
2. Given the rise of competition and the fact that consumers, along with B2B customers, are deluged by these startups screaming for your attention and your money, the money that you would need to pay for hardware inventory is now money required by internet/software startups to buy traffic. This was not true not too many years ago; now the competitive world has changed – the battle for consumer mindshare AND the IT manager’s mindshare in B2B customers has risen exponentially.
3. Some argue that distribution channels for hardware are limited. However, social, word of mouth, and viral mechanisms for internet/software startups do not work any more and you have to market traditionally to build awareness and brand. This more than equalizes the distribution channels for hardware. At least hardware is unique and not just another website product or service which gives hardware a leg up in customers’ eyes simply because of uniqueness. Do we need another photosharing app?
4. Because it takes longer to gain mindshare in today’s world, you must raise for longer runway – 18 months-24 months at least, which means more money is required in any case.
On the hardware side, advantages are emerging or here already:
1. Hardware technology is commoditized and cheap – what was previously rocket science is now readily available in chipsets to everyone. There is not much out there now that requires serious hardware design resources and custom chip fabrication resources. Advanced technologies of the past are now commonplace.
2. Contract manufacturing can make anything on contract – you don’t need your own factories now. When I worked at Apple back in 1990-93, we had our own manufacturing both here in California and in Asia. I designed plastic parts and oversaw the construction of metal tooling to shoot the plastic. We had to build prototypes, test them and their manufacturability, and arrange our own staff to build everything. Now you can hand off all the manufacturing to contract manufacturers, whether here in the US or in Asia.
Because you do not have to setup factories and manufacturing yourself, you don’t have to raise money to do so like in the past. You can raise the same amount of money as your typical internet/software startup and get product in boxes and on shelves.
3. In huge contrast to internet/software startups, there is *virtually no competition*. It is exceedingly rare that when I meet a hardware startup, that I can find another let alone two competitors! This gives hardware startups an unprecedented period of time where they can advance in the face of no competition and grow and learn.
Universities are graduating enormous numbers of software engineers and everyone is racing to internet/software. Hardware engineers, by comparison, don’t exist in nearly as much quantity. In fact, opportunities for hardware engineers are so limited by a wide margin as these skills have moved offshore to Asia. Thus, nowhere near as many hardware startups appear versus the hordes of internet/software startups.
The barrier to entry is experience and knowledge of hardware. Most people fear it because they have never done it; it is natural to avoid that which is unknown.
4. We are now at the edge of what software can do by itself with respect to the physical world. Having humans type in information or self reporting data is just not practical and filled with potential errors. To do better, you need hardware to do the actual touching and sensing of the physical world and connecting that via software to the internet. An example is Quantified Self – self reporting of physical condition has reached its limits; we need 24/7 monitoring of physical condition to get to next level of knowledge, usage, and innovation.
5. Hardware has become small enough, low power enough to do amazing things for long periods of time. Huge possibilities open up due to low power, small size, and accessibility of the technology. Instead of wearing a huge box that is heavy and uncomfortable and needs to be recharged every few hours, we can wear small, unobtrusive sensors all day long, broadcasting vital information to the internet all day and night.
6. By selling hardware, you make money off every sale. By selling software services on top, you further monetize users far beyond the money made from the sale of the hardware. Due to the intense competition, internet/software startups often need to give away services for free which means survival until customers get to a point where they will pay for something you offer, whereas selling hardware means you make money each time you sell it. But then you offer a recurring, monetizable service on top of that to get more revenue.
This is why hardware+software+internet is the key, not just hardware alone.
7. Accelerator programs are now emerging to bring like-minded and experienced people together to enable better hardware product development. Thes are people like PCH International’s accelerator program help new startups get off the ground and provide competitive advantage because the accelerator startups get access to PCH International’s manufacturing capabilities. Other great accelerator programs are HAXLR8R, a combination China and Silicon Valley hardware accelerator, whose purpose is to spend some time in China to learn how to access and manage Asian manufacturing resources.
Other great hardware accelerators are Lemnos Labs, based in San Francisco, and the newly formed Bolt, to be based in Massachusetts.
The declining advantages of internet/software startups and the increasing advantages of hardware+software+internet startups make hardware the next big emerging opportunity.
The evidence is clear. Over the last two years, here is a list of fantastic startups, using hardware+software+internet:
Evoz – advanced baby monitors, analyzing baby cries with recorded data, and giving advice to new parents. [Disclosure: I’m an investor].
Fitbit – activity monitor and scale for better health and fitness.
Lark – silent alarm clock and personal sleep coach.
Zeo – your sleep manager.
Withings – WIFI enabled scale, blood pressure, baby monitoring.
Lumoback – back health and posture tracking.
Green Goose – making everyday things more playful with tiny wireless sensors that automatically measure what you do.
Leapset – a hardware POS play transforming the cash register for local merchants. [Disclosure: I’m an investor].
Metawatch – a platform for wrist based technology development.
Nest – the learning thermostat.
Elacarte – tablet based menu system for enhancing and optimizing ordering for restaurants. [Disclosure: I’m an investor].
And the list goes on. Still, I meet resistance on this issue.
17 years of entrenched thinking, and exploding economies to power profits and return on capital via internet/software startups make the opinion hard to change. And certainly what I argue doesn’t apply to all hardware startups; for example, to build a new car business like Tesla would still require a ton of capital. However, for small, high technology, connected devices this is exceedingly true.
But that doesn’t mean the VC community has to believe what I believe. If there is anything I’ve learned about investing, it’s that the best returns are derived from not following the herd. This is definitely anti-herd, and I’ll either be totally right or I’ll be amazingly wrong – but if I’m right, I hope to be one of the first to ride the wave of this emerging class of startups to success.

Every now and then I’ll meet up with a great entrepreneur with a great idea. But then we take a look at what they’ve built and…gag.
Whatever they’ve put up is pain to my eyes. It’s chaotic, unorganized, a white space hell. The colors are garish, chosen for hexadecimal greatness. The fonts chosen are the best that a browser can offer – comic sans and arial are wonderful aren’t they? All the information is lumped together because stuffing as much as you can into that rectangle called a browser means I don’t ever have to click to get to more information – it’s all there on one page!
The problem is…it’s a visual design nightmare. And it makes me want to run away.
Definitely others will and have. But I sometimes don’t. I’ve bet on the ugly, hoping that the ugly will fix itself later. Sometimes it does, sometimes it doesn’t. But until then, the reality remains. Their site is still ugly. And the problem is, until that is fixed, it will chase others away.
Nobody wants to invest in an ugly startup. It’s a problem on many levels:
1. Ugly startups will chase away customers. Customers are also sensitive to great visual design even if they may not be able to articulate it. Ugly startups look un-professional; there is doubt that this company is for real – if it was, they would find a designer and make it look great, right? They don’t want to look at an ugly startup while using it either – it hurts the eyes. If this is true, then it could stifle your early growth and inhibit trial.
2. Ugly startups aren’t brag worthy. If I invest in you, I want to talk you up. But if someone goes to the site and then comes back to me saying, “Gee that’s the ugliest thing I’ve ever seen” – it affects me. Who wants to hear that? I want to hear, “Wow it’s beautiful and cool – great job!”
3. Ugly startups show a critical team deficiency in design, at a minimum in the visual department and maximally in all areas of design. Designers are the hardest to hire for of any discipline out there; if they don’t have one on staff now, will they ever be able to attract one? Or be able to get ahead of competitors who do have design on staff?
4. If you are ugly and can’t get other investors on board, whether they actually come out and tell you it’s because you’re ugly or not, you’re dead in the water. Other investors will stay away knowing that ugliness lowers the probability that anyone wants to give you money and introduces a higher risk that you’ll die.
Time and time again, investing in startups has been likened to dating. Here is more proof of that – who wants to date someone you don’t find attractive?

It all started back in 2002 when I signed up for my first triathlon and joined Team in Training to help prepare me for it. It also set me on my big science experiment which was “how fast can Dave Shen really go?”
You see, I had a lot of negativity surrounding my first (and also subsequent) triathlon attempt. People told me that my knees would give out, that I better be careful or else I would get hurt. They told me that I was 37 and that trying to race a triathlon “at my age” was a risky proposition for my overall health.
But I didn’t believe them. And now, here I sit 10 years later and still going strong and injury free at 46 years of age. And getting faster each year.
When I got started on my big on-going science experiment, I decided that I was going to stop training with these programs provided by books and magazines. I was determined to train like what I thought a professional athlete trained like – lots of technological and physical support and a great coach too. Someday I should write what it was like to come from zero swim/bike/run experience to completing 6 Ironmans, numerous swim and run races, and ultimately becoming a Total Immersion swim coach last year. For now, I want to talk about one aspect of my training which has led to wiring myself up in many different ways, gaining insight into the future of training and health through technology.
My coach was Mike “M2” McCormack, a popular triathlon coach in the SF Bay area. It was he whom I credit with introducing me to truly data driven scientific, high technology bike training using a Computrainer. We also explored the training via heart rate zones, but we ultimately switched to training via perceived exertion which yielded better results. Logkeeping in Excel kept me on track and now I had a way to go back into my training and look at what went well and what went wrong. I also bought a Garmin 305 GPS watch to track at a detail level my heart rate and performance on runs.
Along the way, I read an article about Andy Potts training for triathlon and Ironman using a data driven, scientific approach. It was a fascinating look into an elite’s training regimen. His coach would take data from his previous workouts and derive workouts for the next day adjusted for his performance on the previous day, and what his condition looked like in the current day! Wow! This meant that if we could gather enough information about our bodies during workouts and our subsequent recovery, we could, in theory, generate an appropriate workout for the day after, whether hard intervals or total rest.
Fast forward to 2011. I was peaking for the LA Marathon and came across Tim Ferriss’s popular book 4 Hour Body. Yet another eye opener on two levels – one was the realization that there is a lot of crap out there on training, health, and fitness, and two, that you have to be data driven in your health and training or else you will never know if you are doing better than the day before.
Going on his suggested diet, I dropped in weight and body fat % to my usual race weight at the LA Marathon, but then I dropped even lower post marathon and plateaued underneath my racing weight! I would never have known that if I had not been tracking my weight and body composition daily. The constant feedback and monitoring were necessary to know that what I was doing had any effect at all, and whether I had to adjust or not. It was gratifying to see that I did not return to my higher pre-race weight.
In following the 4 Hour Body regimen, I got really interested in tracking everything about me. Here is my current complete list of gadgets for tracking me and what I track with them:
Withings Scale – Weight, Fat % sent to website!
Omron Body Composition Monitor – More weight, body fat %, muscle %
Tracking weight, body fat, muscle % for fat loss, and also muscle gain due to weight lifting.
Garmin 305 GPS Watch w/ Heart Rate Strap – the best training tool for running
I store all my runs here, in addition to storing them on Runkeeper. It is good to refer back to my workouts and see what my times are for certain distances.
Finis Swimsense Watch – swim metrics tracking – strokes, lengths
I can track a lot of swim metrics with this watch, but not nearly as many as the ones we need for Total Immersion. Still it’s the best swim watch out there. I now swim with two Swimsense watches to track the action of both arms and am working with the developers on creating a data driven training program for swimming.
Lark Sleep Monitor – tracks your sleep patterns
I don’t have problems sleeping, but I am interested in the amount of sleep I get versus athletic performance. It’s vibration alarm is best in class just as a simple alarm clock.
Fingertip Pulse Oximeter – measure heart rate in the morning
Measuring heart rate every morning can give you insight on whether you’re fully recovered from a previous day’s workout or not. A Pulse Oximeter is much easier to use than holding a heart rate strap onto your chest and reading a watch display.
Microlife Peak Flow Meter – Measures lung function
Following my discovery that had mild exercise induced asthma, I got interested in seeing what my typical air flow numbers are for my lungs. A fancy peak flow meter in my doctor’s office costs several thousand dollars; this little gadget only cost $39.99 on Amazon!
After discovering all these devices, I realized that the cost of these sophisticated instruments had dropped considerably. With the advent of the internet, we are able to take measurements and send them instantly to the Web for storage and further analysis. Most significantly, technology has commoditized a lot of the hardware components required to build these devices. At one time, it took rocket scientists with a lot of infrastructure to create these tools; this is not true any more. And the technology is getting smaller, cheaper, and requires less power than their predecessors. We can now wear these sensors and devices all day and have our measurements beamed to the internet for storage and analysis! Certainly, the Quantified Self movement is gaining momentum where self tracking is promoted and explored.
Still, we are early in the evolution of human data (see my post The Evolutionary Path of Data). Along the data continuum, we have just begun to enable data collection on a regular basis although we are just touching on being able to do it 24/7. For sure we can display/visualize/graph what data we do collect, but it needs to be more complete, consistent and frequent to enable discovery of more knowledge.
Coupled with my experiments in fitness, training, and health, I also realized that the insight we could gain from tracking our body’s metrics constantly was an enormous opportunity. For example, in the area of fitness and training, I see huge potential in guiding people to better health and fitness by exposing the results of what we are doing to ourselves at any time, like what we are eating, how we feel, and how we exercise. Already in Total Immersion, we are experimenting with individualized data driven training. It is my belief that beyond fitness and training, the ability to give us better feedback and guidance on treating our bodies better is untapped and presents the next horizon in health and wellness.
It is gratifying to see startups emerging in this area. Still, we are in very early days on the Quantified Self movement. At the moment, I am resigned to tracking and compiling metrics by hand and then working out insight from the data. I welcome the day when we are uploading our body metrics every second of the day and having intelligent systems tell us where we can do better, stop doing stupid things to our bodies, and live better, healthier lives.