Friday, February 25, 2011

"Launching a web startup became 10x cheaper..."

The conventional wisdom is that starting a web business is (at least) 10x less expensive today than it was 10-15 years ago. It’s said that a decade ago, you needed millions of venture capital in order to launch an Internet startup whereas today, thanks to open-source software, cheap hardware and new ways to acquire users for free (in particular virally via Facebook/Twitter and with SEO via Google), you can do the same with a small fraction of that. And that, while great for entrepreneurs, makes it difficult for large VC funds with many hundred million dollars under management, to deploy their capital because startups ask for less money.

I’ve heard it dozens of times, from VCs, founders, bloggers and others. It’s almost like a mantra, part of the Web 2.0 creation myth, which everyone believes without challenging it. I always wondered if that theory is true, because it didn’t cost Christopher Muenchhoff and me more than about $100 to build and launch in 1997. That, approximately, was the cost of one month of shared hosting, and in the second month, revenues paid for the hosting costs already. The first “big” investment was our own server, around $3000 as far as I remember. We did raise some money to expand the business later, but that was much later – around nine months after launching the service. So it clearly was possible to launch a state-of-the-art web service back then with little to no investment. wasn't the only one, of course, I'm just using it as an example.

When I say “state-of-the-art”, I mean state-of-the-art based on 1997/1998 standards, of course. That's what was needed to be competitive. We of course wouldn’t have been able to build a video streaming site à la YouTube for $100 (hardware and bandwidth was too expensive), and launching an online shop would have been more expensive too (Magento didn’t exist yet). That’s logical, but trivial, and not what the theory wants to say, right?

In other words, I think the theory surely is that “starting a web startup in 2011 that is competitive in 2011 costs 10x less than it cost to build a web startup in 1998 that was competitive by 1998 standards”, right? (If the theory was “starting a web startup in 2011 that is competitive in 2011 costs 10x less than it cost to build a web startup in 1998 that was competitive by 2011 standards” that would of course be true, but I think it wouldn’t mean anything. Being competitive by 2011 standards (often) means that you need an iPhone app. In 1998 there was no iPhone. Get the point?)

And that is what I’m questioning.

Now, if the theory is wrong and it wasn’t that much more expensive to build a web business back then, why did startups raise so much VC at the end of the 1990s? One possible answer is simply “because they could” (and because everyone else did, and you didn’t want to be overtaken by better-funded, faster-growing competitors). Fuelled by a crazy IPO market, there simply was an incredible amount of venture capital available. Maybe that’s the real reason, or at least part of it, why it now appears that launching a web startup was so expensive in the 90s. What do you think?


Anonymous said...

It might not be a question of being 10X cheaper but that the eco-system is 10X richer, all those tools developed and dead ends found make it seem less daunting.

The internet was still to be integrated into everybody's life, the way that its a part of social life now; the opportunities have grown and the internet is now accepted as 'solid' and not some crazy fad.

Perhaps it is an order of magnitude or two easier to contemplate building something on the web 15 years on.

Greg Gentschev said...

You're right that it's certainly not an iron-clad law of nature. It was possible to build a company in the late 90's on open source software and viral customer acquisition for very little. The most notable example that comes to mind is Hotmail.

But overall I think the case still stands. Open source was much less prevalent, both in terms of technology and the talent base. There was no Magento, but neither was there Ruby on Rails, Django, Wordpress, Drupal, Lucene, Hadoop, or countless other projects. Some big, successful sites were built on Perl, but many (most?) companies still ended up on Solaris/Oracle/J2EE. The technology cost a lot more, and it also required an enormous amount of developer time to get stuff built, relatively speaking.

The difference is even bigger in terms of customer acquisition. Remember when companies were paying for $50M sponsorship deals with AOL? Some of that was clearly fueled by drunken stupidity, but there just weren't a lot of scalable marketing models.

A part of the phenomenon that people seem to overlook is that access to information is infinitely better now. With some reading, today's median startup is probably better informed on technology and marketing choices than the top 10% of startups ten years ago. So there are much cheaper resources available, and companies are also much better at taking advantage of them.

Adam said...

I'm getting ready to launch my new startup in the next couple weeks and I've spent under $1000 on everything.

Christoph Janz said...

Good points Greg, thank you for your comment. Yes, I remember those AOL deals – and the billboard advertisings and Super Bowl spots. I think most of these huge marketing spendings probably didn't work out in the end, but it's of course true that if everyone else spends large amounts of money to get the user's attention it's hard to get heard if you're not spending a lot of money too. Arms race... That said, while it's easier to reach millions of users with little to no marketing spendings today, the best startups managed to get a large user base virally or with guerilla marketing even back then.

Spinnakr Active Analytics