The free-download economy is dead

Perhaps you didn't notice, but 2008 marked the last year when SourceForge downloads mattered.

Throughout the year, and for a consistent period of several months, the statistics function on SourceForge stopped working. My own company, which had been tracking our downloads closely, suddenly was flying blindly.

Guess what? Life went on. Not only that, but we moved our central downloads repository to our own Web site because we needed consistency that SourceForge apparently couldn't give us. It also became obvious that download tallies were secondary to something SourceForge never gave us: insight into those downloading our software.

The download economy, in short, gave way to the customer economy, at least for Alfresco.

Chris Anderson, longtime champion of "the Long Tail" and the power of giving stuff away, has captured this idea well in a Saturday column in The Wall Street Journal. He argues that the allure of "free" is no longer as strong as it once was for businesses that need to make a buck:

Until September of last year, the model (for companies trying to build a business on the Web) was pretty simple. 1.) Have a great idea. 2.) Raise money to bring it to market, ideally free to reach the largest-possible market. 3.) If it proves popular, raise more money to scale it up. 4.) Repeat until you're bought by a bigger company.

Now steps 2 through 4 are no longer available. So Web start-ups are having to do the unthinkable: come up with a business model that brings in real money while they're still young.

In boom times, it's enough to be popular. In down times, you deservedly die if you don't make money. Open source has shifted from the "free" economy to the customer economy, and downloads have largely become irrelevant.

Of course, open-source companies still need to drive downloads. Downloads feed into leads, which feed into customer conversions.

The critical strategy for open-source vendors is no longer to find ways to goose download numbers; it is to devise product strategies that encourage more paid adoption, something I addressed at OScon in 2006 that has become much more important in the past two years.

This is why Sun Microsystems' MySQL has been changing its business model to accommodate commercial add-on services that help drive paid subscriptions to an otherwise free database. It's why Red Hat has always constrained access to its Linux binary distribution. And it's also why Canonical's Mark Shuttleworth is going to have to move beyond "free" to create a compelling business out of Ubuntu.

SourceForge download statistics didn't work for much of 2008, which helped us to see that they largely don't matter, anyway. Conversions matter. Customers matter. Welcome to the open-source customer economy.

Follow me on Twitter at mjasay.

[PBarila]

Matt, In your previous post about MySQL changing its business model, you argued that they should differentiate the database versions, and Zack replied and said they were committed to making sure the Community Edition doesn't become "crippleware". Sounds like maybe you might be starting to agree?

[shootthecops]

It sounds like the issue is the bandwidth costs that prevent places like SourceForge from hosting properly for free. The U.S.'s slow, archaic internet service no doubt is the culprit. If I had a company that suffered at the hands of U.S.'s pitiful speeds, I would get outta dodge and set up shop in greener pastures. The U.S. has embarrassed itself by letting its own connections become so slow and businesses will be the ones to pay for not investing into making the connections faster and cheaper.

[arjenlentz]

Horay for real business models, I've long been advocating this.
On some of the examples you used I think the jury is still out on effectiveness/sustainability.
Also, companies don't just need customers or sales, they need mindshare. They need to not just be seen but really have to be "good" and "nice".

[swag]

It's definitely a market with low risk, low reward, and, to be honest, low ambition right now. Companies that do get funding seem to be building what I would merely call a 3-month feature a couple years ago.

As for MySQL, they are in for a rude awakening with v6 pricing when they see how much Oracle has laid the groundwork to undercut a lot of it. It won't be pretty.

[fazalmajid]

The notion that Oracle will undercut MySQL pricing is so ridiculous it doesn't deserve a rejoinder. Oracle is a company that hikes it prices just because it can, without even the pretence of introducing a new release the way Microsoft does.

Oracle's game plan with MySQL is crystal-clear - they bought InnoDB so they have their hand on MySQL's jugular. that's why MySQL is trying desperately to develop the Falcon replacement for InnoDB, but that project is floundering.

The MySQL freemium model is a viable one as long as they don't violate the trust of their users by attempting to cripple the base product. Unlike Oracle, they can't just coast on the lock-in from their installed base, however, and will have to earn each dollar in services or add-ons by delivering value.