daily articles for founders

Are startups 10x cheaper than ten years ago?

This article was posted last week, and spurred an interesting discussion on Hacker News. At key is the question of whether startups are really cheaper than they used to be.

This is, of course, relevant because the idea that startups are cheaper than ever to launch is often brandied about to support this or that argument, whether it's about the increased number of tech startups, or the investment climate, or the reasoning about why you should start a startup right away, or the disruption to the VC industry, and so on.

Thesis and antithesis

One side of the debate claims that actually, startups are no cheaper than they used to be. In the late 1990s, it was possible to launch a startup on the cheap, on a shared hosting package, for puny amounts of money, and wait until it picked up before investing in more expensive infrastructure (just like it is today).

Hiring has not gotten cheaper, and infrastructure costs on the low end are just as low as they used to be, and so, the argument goes, it isn't cheaper to start up today than it was ten years ago.

The other side of the debate points out that many startups in the late 90s spent a lot of money buying overpriced Oracle licences, hiring rare and expensive specialists to deal with all this pricy technology, and that this was part of the cost of doing business. Today, MySQL (free) has replaced Oracle, and there are a lot of free hosting packages, and there are hyper-productive technologies like Ruby on Rails which allow you to build an entire web business with a handful of people and dollars.

Hiring costs may not be cheaper, this side claims, but technologies are more productive and deliver better results.

To this, the first side replies that you didn't need to buy Oracle in the 90s any more than you do now, and that this was just an inflated cost caused by too much VC investment in unsound startups. And, ten years ago, a tiny team of hotshot programmer could build a business out of nothing just as well as they can today.

And so the endless Saṃsāra of arguments goes on. If you have a few hours to kill, you may wish to join in!

The reality

As with every unsolvable debate, the truth lies... on both sides.

It was indeed just as possible to start a startup on the cheap ten years ago as it is today. Many people did then, and many people do today. If you're scrappy, smart and determined, you can start a tech startup for pennies. That's as true now as it was then.

But not all startups are started at the very cheapest end. Some businesses do require investment of some sort before they're viable businesses, or even require it as growth money. And this amount of money has been going down, at least as far as the technical expenses is concerned, because both the technologies (Rails as well as others), the infrastructure (AWS, Heroku, but also the vast number of supporting apps that make it easier to do things like accounting, support, or user testing more efficiently), and even the available knowledge and mentoring (via movements like the Lean Startup, or even communities like Hacker News), have been getting better.

AWS may not be the cheapest, but as a "standard option" for the scale-minded startup, it's a damn sight cheaper than spending thousands of dollars a month for a bunch of dedicated Rackspace servers, or buying up colocation space, physical servers and sys-admin time for a similar amount, as you might have in 2001.

It is possible to get further with less money than it was ten years ago.

And that's the crux of the matter. As with Parkinson's Law and time budgets, startups will expand to take up the funding available. But they will get much further with the same amount of money than they typically did ten years ago. Of course it was possible to start up on the cheap ten years ago. But:

  • today's cheap startup will get further than yesterday's cheap startup;
  • today's mid-range startup will get further than yesterday's mid-range startup;
  • today's expensive startup will get further than yesterday's expensive startup.

Startup opportunities

Perhaps more importantly, there's another factor at play: the opportunity-space.

Thanks to the blossoming of internet users, countless web technologies, open APIs, and other web companies, there are more opportunities to build startups today than there were ten years ago.

This seems counter-intuitive at first - surely, back then there were so many more things left to do. If you think about it, however, you will see that many opportunities depend on other things being in place. You couldn't have started Twitter or Groupon in 2001. Not enough people would have understood either of these platforms. You couldn't have started any of the companies, like Instagram, Foursquare, or Bump, that thrive on the iOS ecosystem - it didn't exist. You couldn't have started DropBox, Heroku or Mint back in 2001 - the infrastructure wasn't there to support them.

The fact that there are so many opportunities to build something useful and monetisable is, in my opinion, a bigger factor in the rise of web startups than how cheaply a scrappy founder can launch a product.

And as far as this particular trend is concerned, I'm pretty certain we are still at the beginning of the tsunami. Cheap or not, expect many more startups in the next ten years than in the last.

Update: Someone pointed out that I didn't answer the question in the title, which is clearly very bad form and I apologise for this faux-pas. The proper answer to that question is: you can't compare startups directly. What you can do with a certain amount of money depends directly on when you start doing it, and what you're trying to do.

What you can reasonably say is that the scope of what you could do 10 years ago with $X is about the same as what you can do today with $X/N, where N is larger than 1 (but perhaps not as big as 10), because of the improved opportunities, technologies, and supporting infrastructure.

More from the library:
Product/market fit vs profitability
Patrick McKenzie's work on the Fog Creek website
Always drill down to fundamental metrics
Google Analytics Alternative