Want to improve your odds? Think hardware

I think the general idea is very good.

Hardware is "hard", but it's hard for the competition too - if you're good in that area that should be what you want. The social network / web app market has few barriers to entry - the programming is not difficult enough to prevent average programmers with great marketing and design skills from entering. Not good if your strongest asset is your technical ability and you hope go head-to-head with web 2.0 hipsters on there terms.

If you can handle something that requires math, physics, hardware knowledge, big initial investment, etc, why not look into ways of monetizing it? It's tough but you'd have a smaller pool of competitors than if you decided to try and make the next Flcker/Facebook/Twitter/etc. That more people aren't doing this leads me to believe I'm overlooking something, but it's surprising so many geeks are drawn to (and succeeding at) startups that seem to be the antithesis of geekiness (social networks, advertising, etc).

Hardware is ... HARD. A lot harder than software. There are no nice frameworks assuring your board works or that you aren't paying five times the cost on your plastic parts because of poor mold design. Iterations take months not days. This isn't something you can bootstrap with 4 people, the first plane ticket to China to meet with contract manufacturers is going to cost more than the entire server running a web app for 2 months. The container shipping it back will cost more than the food two founders eat a web company in a year.

BUT - at the end of the day you have a real product, which people will expect to pay for, and a barrier to entry behind you.

I've been thinking about this a lot lately. It seems that hardware may be the New Way (tm). There exist some pretty exciting technologies for cheap, namely FPGA boards and SOC (system on a chip... and/or nano-itx). These make a very large portion of hardware design much simpler. We can suddenly apply software rapid development techniques to board design with the FPGAs. Further the SOC makes much functionality within a product just software development.

Further, there are many exciting "hobby kits" which make gadgets very creatable and testable, since they allow for easy plug-and-play hardware creation. This sort of thing makes prototyping very very easy, and the rest can be outsourced pretty well. Sure there are complications that I know nothing of, but I think there is an abundance of supply chain for a lot of projects, making small hardware companies totally possible.

A slightly different perspective on the above: I beleive there exists at least one market that can be exploited by putting together existing big components in fun new ways.

Ugh!

I can't imagine a worse market for a small starup than consumer electronics. Margins are tiny, getting distribution is insanely hard and you need incredible amounts of cash to start up. And if your product is popular, it's a matter of weeks before the guy overseas building it for you copies your design and sells it himself for 1/5 the price.

Far better to find a small industrial niche and enter it with a simple product the "big guys" won't bother with. You can sell directly to your users, your margins can be much greater, with smaller volumes you can literally build the hardware in your basement and your investment is minimal.

My sales are low (less than $10,000 last year), but by targetting industry, I get >$100 for essentially the same product that is marketed to hobbyists for $20.

That's an interesting article, although I don't particularly agree with the conclusion.

As an aside, I think "No matter how hefty the investment, the odds of success are high" is a typo. In its context, this would seem to go against the author's main point (namely, that opportunities are better on the hardware side of things). But the author is there speaking of software... and aren't high odds good odds?

The article seems fixated on consumer electronics. But in the same way that software is built and sold to businesses, the same can be done with hardware. Not everything has to ship in a shrink-wrapped box.

I don't know much about that market, but I'm very interested in learning more.

Hardware companies have to handle inventory and RMA. Both are potentially 7-figure annual expenditures and you can't just build them into your price year 1 like you do with shipping.

Has the rise of virtual circuit design helped to improve margins for the hardware industry? the ability to test a prototype before building it must count for something.

I have a really hard time believing that the average margin on consumer hardware is 30%. Very few companies, with the exception of Apple, can pull off that kind of margin. As a new market entrant, you're going to have to significantly discount your hardware to attract consumers. Even established players, like Microsoft with the Xbox, sell hardware at or below cost to sell software.

pg also suggested this, http://ycombinator.com/ideas.html (number 27)

Given how hard I found it to find a decent alarm clock that didn't totally suck in terms of usability, I agree. I've also had 3 digital TV receivers that all were atrocious too.

Only word of warning would be to read Alan Cooper's inmates are running the asylum first, he has some spot on observations about how modern electronics are treated like computers (e.g. ovens that are hard to use, even though the old analog design, a dial with markings on, worked)