First and Second-order Effects on Market Structure and Consumer Choices
In this blog post I will discuss two mechanisms of network effect and how they affect market structure and consumer choices. The content is based on the chapter Personal Data and Collective Value: Data-driven personalisation as network effect in the book Data-Driven Personalisation in Markets, Politics and Law written by Nick O’Donovan (link).
O’Donovan describes two types of network effects: first-order and second-order. First-order effect is what we learned in class, where if more people (nodes) are using a certain product, I am more likely to use it too. O’Donovan characterizes this effect as quantitative, since it depends on the number of connections (edges) or number of people joining the network.
On the other hand, second-order effect is a qualitative effect, and I will cover this in more detail as it was not covered in class. The qualitative effect is a more modern effect facilitated by big data methodologies, where the more data a company has about users, the better quality service it can provide, thus enticing more users to join. For example, Netflix can provide excellent suggestions based on its abundant user data. A pure second-order effect means I’m joining Netflix because it provides better viewing suggestions, not because my friends are on it.
From class we saw that price is a lever we can pull to move the adoption tipping point. While price is very effective with first order effects, it is not as effective with second order effects. Second order effects are not only based on market share, but also on length of relationship. For example, Netflix has historical streaming data on users spanning over 10 years. Even if a new streaming company gained the same percentage of market share as Netflix (same first order effect) by price reduction, it cannot mimic the second-order advantage that Netflix has. One could argue that data sharing can mitigate second-order inequality, but this comes with a host of data privacy issues. Even if companies wanted to share anonymized data with newcomers to the market, significant privacy issues would prevent this from happening. In fact Netflix published its data for a recommendation competition, but privacy issues forced the company to pull the competition. The downside of the second-order effect is that the “first mover” advantage stifles new competition, resulting in fewer choices for consumers.