How To Prioritize Network Effects For Your Product
In early 2018, James Currier identified the 13 types of network effects that the NFX team has uncovered so far as listed below:
This is an essential foundation for anyone building network effect businesses. In this post, you’ll learn how these network effects can be classified and use that information to aid with prioritizing them.
Classifying Network Effects
Before we classify network effects, we need to understand the architectural framework of network effects products. Sangeet Paul Choudary visualized this framework as a “stack” with three layers:
a. Interactions Layer where users explicitly or implicitly exchange or create something of value with each other. Some of these interactions might even happen outside the product. This is the part of the product that is publicly identifiable as as the network, marketplace or community.(e.g exchanging messages, rating app developers).
b. Technology Infrastructure Layer that enables value creation. This is what the producer/supply side creates its unit of value exchange on. (e.g app developers, eBay sellers)
c. Data Layer that is produced as a function of value exchange. This in turn is used to increase the relevance and ongoing utility of the product. (eg Google searches, Yelp ratings).
It follows that the 13 network effects are a function of one or more of these three layers. If we map these two concepts to visualize their correlations we get:
Depending on the core value of your product some network effects may not be available to you (at all or to start). You need a way to prioritize the ones you could deploy as this will impact your initial product considerations.
The answer to the prioritization question lies in your initial traction model. Tactically, there are many ways to solve the chicken and egg problem that network effect businesses face. Like with network effects, these tactics too, lend themselves to being classified based on the kind of product you have.
You have two choices when it comes to building your product:
- Connection-first: Get users first. The act of connecting them leads to content, conversations and transactions
- Content-first: Start off as a standalone product that provides utility. The data or content created because of this use enables connections and transactions between individual users
You can read more on these two models in this post by Sangeet Paul Choudary.
This decision is critical towards understanding whether your initial network effects will be a result of:
- interactions between users
- what the product enables producers/suppliers to build
- the product itself
If you’re going down the Connection-first route, your initial success depends on user interactions. Over time, the technology can become more robust as these interactions give you information on what users find valuable. This in turn will allow you to also exert data- enabled network effects to make the product more useful.
In a Connection-first product, order of network effects from Table 1 will look like this:
If you opt for the Content-first route, you are choosing “single player utility”. You don’t need to start with user interactions. But it means you need the technology infrastructure to enable value (content) creation and capture data to make the product more useful.
In a Content-first product, order of network effects from Table 1 will look something like this:
Depending on the core value of your product, your initial focus may be more on network effects due to the underlying technology that enables value creation or on data generated from usage. Irrespective of the order, you will need both of these before you pull the lever on user interaction related network effects.
Use this template from the NFX team to first identify which network effects your product can have. Then use the prioritization methodology from this post to help you decide which ones to focus on when building your product.
What was your biggest takeaway? Leave me a comment below.