Network effect, data network effect and their implications
In class, we talk that the some of these products have network effect. As more people use them and they have more values, even more people will willing to pay the price for them, and vice versa. Classic examples can be Facebook and cell-phones. In this blog, the author introduce a specific kind of network effect–“data network effect.” Unlike Facebook or other communication tools that gain network value by connecting people into its network, other services and products like google search and Amazon’s recommendation system aggregated users’ data to estimate what the user are looking for. These services usually employ machine learning methods to train itself. Recommendation systems cannot work well when the data is limited, and if it works well and provide accurate recommendations, users more likely to purchase more items and they can gain more data. Google search does the similiar thing. This 2008 article proposes that Google harnests both standard network effect(which it calls as direct) and data network effect(which it calls as indirect).
The standard network effects we discussed in class may be strenthened when work together with other network features. For instance, the network effects suggest that marketing is crucial in deciding whether some start-ups will prosper or die out, and the success of marketing is also influenced by “richer gets richer” effects and information cascade. For online marketing, if their pages is influential at the beginning, they are more likely attract more in-links, which means more people might see the link towards their marketing pages. Also, when someone see many others using a product or just hearing many others saying something is popular, they are likely to think that thing is”popular.” The impression that something is popular spreads in information cascade, and can become powerful marketings.
By the way, I read someone’s post about the spread of driverless cars and cluster density. I am thinking that the application of driverless car can be an extreme example of network effects. The saftey of the driverless car still poses challenges to sensors and algorithms, but if all human driving cars are replace by driverless cars and some network connect them together, it seems that they can be safer and the transportation may also be more efficient as all cars can communicate with each other.