Monthly Archives: April 2011

Recommendation System

This is a paper on the Netflix Prize. Someone needs to build a personal trending topics for Twitter. I follow close to 100 people on twitter, and maybe 150 if you also count lists. Each day, I receive hundreds of tweets. I need to be able to identify quickly 20 tweets that are the most interesting to me.



Untangling Advertising: An Analysis of the Basics

A critical component of today’s technology economics is advertising. Like it or not, we as a society put a lot of effort and intelligence into getting internet users see ads, click on them, and potentially buy the advertised products. It’s important to have an understanding of the industry and see where it’s going.

Most of the major tech companies derive some revenues from advertising, ranging of little in the case of Apple and a lot in the case of Google. In fact, a major part of Google’s revenues come from its ad products, Adwords and Adsense.

Let’s look at some basics of online advertising. The economics of advertising are the same online or offline. There are generally 3 primary parties, the advertisers, or the companies who want to increase their products

Asymmetry of Motivation, Ability

Many times, when people are explained a start-up concept, they would immediately say that the concept has no chance because some established companies could do it easily. In technology, it’s Google. “Oh, Google can squash that,” or “What if Google decides to do it?” they would say. In reality, many start-ups succeed in the presence of hungry big companies. It’s also true that many start-ups just get washed aside by the big competitors. So how would we know which has a chance and which doesn’t?

In his book, “Seeing What’s Next,” Christensen argues that in order for an up-start to win, they must have either or both of Asymmetry of Motivation or Asymmetry of Ability. A big company like Google, due to is “value,” must focus on big opportunities (it’s social for Google right now); therefore, it will not have the motivation to go for smaller markets. Examples of this asymmetry are some smaller SAAS areas such as data warehouse/reporting and SEO. If start-ups in these areas succeed, they can be acquired by Google but not overthrow it.

Asymmetry of ability is a bigger concern for big companies. Facebook and Twitter, for examples, have acquired abilities and asset (the users, bounded by the network effect) that allow them to survive Google, even though Google has great motivation to succeed in social. Google Buzz did garner some buzz at first, but that has cooled down significantly and it may have become irrelevant. On the other hand, I believe Color, a great idea in localization, does not have either asymmetry of motivation (both social and localization are important to Google) or asymmetry of ability (I think data analysis and creating working algorithms are Google’s strengths, and a start-up cannot overcome this). So unless Color is able to erect some sort of barrier through the network effect, it may just be swallowed up by a company like Google soon.

Google is working hard on social, which should leave a bunch of smaller opportunities available for start-ups.

Outstanding Analysis with Prediction

I came across this great analysis, succinctly written by Rajan. I like it because it demonstrates that everything is interconnected, and if you are able to see the important interconnections, you can see how the “system” functions, which allows you to make some good predictions (or conversely, it allows you to assess more accurately the causes of what’s happening right now).

When we look at Paul Ryan’s recent proposal through this lens, we’ll see that his proposal is more opportunistic than courageous. He made the proposal because his constituents’ sentiments had shifted toward an extreme during the debate, and his proposal pocketed him some short-term political gains while not truly addressing long-term issues such as the tax cuts for the rich. That is not courage; political courage to me means that you are willing to give up some self-interest short-term gains for the good of the country. Getting wild praises from your base (and ┬ábeing reviled from the opposite side) does not make you courageous, it makes you politically smart. So please people, give him credits for his political savvy, but not courage.

What’s The Future of Internet Coupon?

It’s written many times about the wide range of experiences restaurants go through in dealing with the internet coupon craze. Groupon is currently the most recognized player, but how will this game end up?

(to be continued)

Smart Companies Make Egregious Mistakes: Cisco and Pure Digital

It’s fascinating to see companies with good track records make some egregious mistakes. Cisco bought Pure Digital, the maker of Flip Camera, just 2 years ago for $590M. When Cisco made the acquisition, the iPhone 2 was about to debut. While it didn’t yet have a video recording function, that functionality is inevitable. Two years are not very far out to predict roughly how markets will turn; the demise of Flip was already in motion at the time of the purchase.

When this type of mistakes happen, you have to wonder what the internal dynamics were like for them to make such a bad decision. There are several hypothesis: incompetent managers, disruptive technology, poor group dynamics. I tend to dismiss the first reason for companies with good track records such as Cisco. While you can argue that the lower end camcorders in smart phones disrupt the consumer video cameras market, this reason doesn’t explain Cisco’s mistake because the disruption is already taking place at the time. This leaves us with some issues with group dynamics that lead to poor decisions.

3D Printing: A Disruptive Force

As 3D printing gets cheaper, it will disrupt many markets that have long depended on scale as a key part of their business models. The assembly line manufacturing method relies on producing same products in order to lower costs. This traditional manufacturing method narrows many markets for a couple of reasons. First, testing out a non-software product is very risky and costly because a small order for a test product does not have scale and would be priced much higher than would an established product; therefore, many potentially great concepts will not even see the market. Second, because in many markets prices greatly affect demands, companies have tried to lower product prices by producing a small number of products that roughly meet the needs of an entire market rather than producing many differentiated products that closely meet the needs of sub market segments.

With 3D printing, these compromises are no longer necessary because the unit cost for a batch of one is not much different from that for a batch of many. Because of this new found freedom from scale, it’s now possible to produce differentiated products for niches of customers as well as testing out new innovative products. Mass production is now replaced by mass customization. How will this drastic shift disrupt the current system?

First, we need to have a more accurate assessment of the shift. I don’t believe that economy of scale will disappear. At the factory, it’ll still exist, not at the product/model level but at the total production level, i.e., a big factory would still have a lower unit cost than a smaller one, at least due to its market power. At the point of consumption, I am sure it’s still a little costlier to print at home than it is at a factory, but the real comparison is between the cost of printing at home vs. the cost of printing at the factory plus other costs of getting the products to the consumers, including delayed gratification cost.

(to be continued)