It’s so easy to start a company, everyone’s doing it now

To paraphrase investor Peter Thiel, we were promised flying cars, and what we got were 140 characters.

But while I’ve often felt that way myself, I think that complaint is a little too facile.

Thiel’s quip echoes a common lament among those, like me, who were drawn to Silicon Valley because of its legendary achievements: the transistor, the integrated circuit, the Internet, and the personal computer. (Well, the PC actually came from Florida, but it got out of that state ASAP, and it’s fair to say that Silicon Valley companies like HP and Apple had an enormous amount to do with the PC’s success.)

These were life-changing, even world-changing, inventions. So it’s a bit disappointing to survey the landscape of innovation today and see the brightest minds of a generation chasing comparison-shopping solutions, social media aggregators, and subscription commerce schemes for selling razors, condoms, and underwear.

In fact, it’s easy to overlook the tremendous chain of invention that made those 140 characters possible. Twitter couldn’t exist without cellular data networks, smartphones, widely accepted browser standards, and widespread Internet service. If Twitter solves a relatively simple problem, it can only do so because it rests on the shoulders of giants who were solving incredibly difficult problems.

Ditto for the companies that we love to mock because it seems their hardest thinking has gone into deciding what kind of swag they should send to tech bloggers. You shouldn’t overlook that these companies are possible only because they stand at the end of an incredibly long line of inventions that have made it cheaper and easier to start a company than ever before.

Aaron Emigh, the founder and chief executive of Shopkick and a serial entrepreneur, drove this point home for me recently.

A couple years ago, he compared the costs of building an e-commerce company to scale (the point at which it can support millions of customers) in 1999 versus the costs for building a similar company in 2009. He helped build one such company in 1999, CommerceFlow, so he had real data on which to base his analysis.

What Emigh found is that bandwidth costs declined an average of 40 percent per year from 1999 to 2009. Internet capacity that would have cost a startup $180,000 per month in 1999 would only cost $6,000 per month in 2009. That’s a huge difference.

Bandwidth, of course, is cheaper because of all the fiber that got laid during the dot-com days. People used to scratch their heads about the “dark fiber” that had been built by companies like Qwest and Global Crossing. But the truth is, that dark fiber (unused optical cables for Internet data) is now cheap fiber, and it’s in use. So maybe the dot-com boom was good for something after all.

Bandwidth is only part of the story, though. Cloud services mean that fledgling companies can get started on little more than a founder’s credit card and can grow to a pretty respectable size without incurring significant capital expenditures (capex) at all.

In 1999, Emigh spent about $10 million buying expensive Sun servers and other hardware necessary for hosting a high-volume e-commerce site.

“These were really expensive machines and you needed a bunch of them,” Emigh said.

In 2009? Total capex dropped to just $27,000. It could probably be close to zero today, as all you’d need to buy would be desks and chairs for your few employees and maybe a few laptops.

Overall, the total monthly expenditures needed in Emigh’s model dropped from $479,000 to $162,000, amounting to a 66 percent discount.

We’ve seen other changes besides costs. Distribution is easier thanks to wide-reaching search engines like Google, and advertising is cheaper and more effective.

For mobile startups, Apple’s and Google’s app stores give developers access to tens if not hundreds of millions of potential customer phones, a feat that was next to impossible to achieve 10 years ago.

Engineering productivity has gone up, too, Emigh told me, thanks to the maturity of open-source tools. A development project that would have required 20 engineers in 1999 might only need seven engineers now.

And we know much more about how to start a company. I started a company myself in 1999 and, it is fair to say, I had almost no clue what I was doing. Today, incubators and seed funds like 500 Startups, Y Combinator, and TechStars provide systematic coaching and guidance. The “lean startup” philosophy has become widespread through books, websites, and even giant robot dinosaurs so that even outside of an incubator, a would-be entrepreneur can learn a lot about what it takes to get started.

So, yes, we still don’t have flying cars. What we got were 140 characters and a host of Pinterest clones. But let’s not forget that those Pinterest clones depend on the engineering and business equivalent of a flying car: a cheap, ubiquitous, and robust infrastructure for building companies more easily than ever before.

Photo credit: x-ray delta one via photopin cc