The Gong Show

Anand Rajaraman, who is teaching a data mining class at Stanford, wrote up a great example of the power of a superior data asset. Anand instructed his data mining students to break into teams and create entries for the Netflix prize. Here’s what happened:

Different student teams in my class adopted different approaches to the problem, using both published algorithms and novel ideas. Of these, the results from two of the teams illustrate a broader point. Team A came up with a very sophisticated algorithm using the Netflix data. Team B used a very simple algorithm, but they added in additional data beyond the Netflix set: information about movie genres from the Internet Movie Database (IMDB). Guess which team did better?

Team B got much better results, close to the best results on the Netflix leaderboard!

We, at USV, talk about the advantage of more (or better… or proprietary…) data all the time. Brad wrote about the subject in his December post on Google’s data asset:

Google has so much more data at their fingertips that even if a startup does a much better job leveraging data to deliver recommendations, Google could potentially provide a better value proposition to the end user with an inferior algorithm powered by more data, sourced from a broader range of services.

Brad’s example makes an important point: data is one of the few remaining means of defensibility. In the first dot-com boom, you could find defensibility in patents, out-fundraising your competition, proprietary code and algorithms. Now, by contrast, no one respects patents (and they’re too costly to defend), web services are so capital efficient that out-fundraising your competition is just a distraction, and open source code has eroded the advantage of proprietary code and algorithms. The main source of defensibility that remains is in your data asset. If you can aggregate more data, license more proprietary data, generate more of your own implicit usage data, or crowdsource more data than your competitors then you will be at a significant and defensible advantage.

James Carville hung a sign in the Clinton campaign headquarters in 1992 that said, “It’s the Economy, Stupid” as a constant reminder of what fundamentally mattered in the process of unseating George H. W. Bush. I would love to see some pictures of people with It’s the Data, Stupid signs in their startup’s office.

I think my mixtape (on Muxtape) is done… I say that hesitantly because I know I’ll fuss over it some more over the next few days, but I think it’s generally in its final resting place right now. Enjoy!

Photo Credit: Lee Jordan

Albert Wenger writes about the structural changes in both the firm and in web services that make a B2B marketplace more viable today than in the original dot-com boom. It’s a thoughtful piece that’s worth a careful read, and I want to add my $0.02 here:

Albert surfaces some remarkably interesting facts about how non-employer businesses (business where the owner is the only employee) are growing rapidly.

For instance, the US Census Bureau tracks non-employer businesses, which are businesses with no paid employees other than the business owner(s). The number of such businesses increased from 15.4 million in 1997 to 17 million in 2001 or about 10% in 4 years, but then grew by 20% over the next 4 years to 20.4 million in 2005 (most recent year for which data is available).

I think this shift in the makeup (or deterioration) of firms will have many interesting impacts. For example:

• We will see an increase in micro-acquisitions for the primary purpose of building HR. Large firms will buy one-person companies instead of trying to hire people. This is already happening today: RockYou and Slide are “acquiring” Stanford kids from the Facebook Apps class that BJ Fogg and Dave McClure ran. Instead of trying to hire these grads, RockYou and Slide are giving these kids an acquisition price that is the functional equivalent of a large starting bonus.
• The SMB services market will become very competitive, and specific verticals will commoditize as this market balloons and entrepreneurs chase the gold rush.
• The blending of consumer and enterprise software will increase rapidly. When the difference between a firm and a consumer is blurry, the software targeting a non-employer business will be equally as blurry in terms of its defined target user. For example, AOL Instant Messenger and Salesforce.com will start seeing increasing overlap in their user base.

In short, this growing market of Non-Employer Businesses is ripe with investment opportunities, and they will look very different from the current Enterprise and SMB investment opportunities. I’m excited to chase it down.

Here’s an interesting comment on software licensing from a recent article in Wired regarding the business of open source software:

“I think the software-license business model is archaic,” says Kevin Harvey, a venture capitalist at Benchmark Capital, which recently cashed in on its investments in MySQL and the open source mail-client firm Zimbra, which Yahoo picked up in late 2007 for $350 million. “I wouldn’t fund a company with that model, and I don’t think anyone else would, either.”

I definitely agree with Kevin’s statement, and, in general, I’m not interested in software businesses that employ a straight pay-for-license model. But, that means I (and Kevin) am being dismissive of a $270bn business: Microsoft. As fun as it is to proclaim that Microsoft is Dead, MSFT has an army of talented developers, a mountain of cash, and a couple monopolies that aren’t going anywhere anytime soon. Will they evolve away from their software licensing revenue? Only if they are forced to do so by disruptive competitors in each of their software verticals, and in the mean time they’re going to continue to mint money using an “archaic” model.

I have to reiterate that, especially in the long run, I 100% agree with Kevin on this issue, but it’s hard for me to reconcile my opinion with Microsoft’s success.

@msg went around and did short video interviews at SXSW asking people a simple question: WTF is Twitter? As one might expect (considering how amorphous Twitter is), the responses are all over the map. Go check out the project and @msg’s breakdown of each contribution for a deeper dive.

My favorite responses are @innonate’s “EKG” and @laughingsquid’s “people mover,” both of which are embedded below.

Twitter according to Scott Beale from michael galpert on Vimeo.

Twitter according to Nate Westheimer from michael galpert on Vimeo.

A simple open question: what web services do you pay for online?

My answers are:
- Flickr: I have a pro account… $24.95/year.
- eBay: I have sold items on eBay, thus I pay them listing fees and a % of final sale. This has only been a couple dollars over many years of eBay usage.
- Akismet: This is the anti-comment-spam service from Automattic. Based on my level of usage, I don’t have to pay for it, but I do anyway. I’ll likely stop paying for it now that I have switched to Disqus. I believe it was $5/month.
- Dreamhost: They host my blog, in addition to a number of sites and email accounts I manage in a consulting gig for a little income on the side. It costs about $10/month.
- NetFlix: I consider this a web service because the majority of my NetFlix consumption comes through their new online streaming offerings. Cost is $13.99/month.
- Skype: I don’t have a Skype Pro account, but I do pay for SkypeOut minutes. It’s a heck of a lot cheaper than getting a landline. I spend about $3/month in SkypeOut calls.

And, that’s all. It’s amazing how much utility I get from web services considering how little I pay in cash. Of course, having seen hundreds of web services’ business models, I know that I pay for the majority of services I use in two ways: 1) my attention to advertising and 2) the data I create through my usage of a service.

Seems like a great deal to me.

In this episode of MobLogic.tv one of the interviewees says that due process in the law should take a backseat to “going out to find bad guys and kill them.”

This is a small example that represents why I dislike the phrase “the bad guys” when used in a logical (often political) argument. “The bad guys” is an allusion to the black-and-white nature of comic book characters, like Superman (the epitome of all that is good and moral) battling against Lex Luthor (the irrationally evil bad guy). In general, I prefer Japanese comics to American comics, because Japanese comics rarely use the stark “good vs bad” plot device.

Though it would be easier to believe otherwise, the bad guys don’t exist in real life. When someone resorts to the “the bad guys” generalization, it is a failure to comprehend other peoples’ motivations.

I had an interesting conversation recently with an analyst candidate at USV about the transport of bits (ie zeros and ones, coded information) vs atoms (ie physical goods, people, objects). It seems that the transfer of bits is largely a solved problem. But, we’re still figuring out the transfer of atoms…

Some big companies have been built by introducing innovative solutions to increasing the efficiency with which we can transport atoms. FedEx is a great example.

Other big companies have been built by replacing atoms with bits. For example, prior to PayPal, eBay users were mailing checks or money orders to each other. But, then PayPal figured out how to email money… they changed the dynamics of the problem from a transfer of atoms to a transfer of bits.

I think Amazon’s recent announcement regarding fulfillment services is big. It’s more FedEx than PayPal, and it’s another step in the evolution of transferring atoms. I’m excited to see what the implications will be for the long tail of retailers.

After an attempt to use Intense Debate on this blog, I reverted back to the original comment system with my blog. But, recently, I’ve been getting some complaints about the AJAX in my comment submission system breaking. So, now I’m trying out Disqus. Let me know if you have a notable (positive or negative) experience with it.