This post will look at the network effect, the most critical concept in the digital world, and startups. Throughout this article, we will first look at some of the histories of the formation of this concept. Next, we will discuss why it is crucial to becoming familiar with this concept. Then we try to understand the difference with concepts close to it, such as viral growth. Next, we examine the different types of networks, their architectures, and their impact on our product strategy. And finally, we get acquainted with different kinds of network effects.

Does reading these posts benefit me?

Before you start, take a look at the resources section. If you have not seen them before and are not familiar with them, and one or more of the following sentences is a good description of you, it will probably be helpful for you to read this article. (At least I tried to make it so)

• In a startup, I’m a product manager.
• I am not a product manager now, but I make decisions directly in a startup’s design and product development planning.
• I would like my career path to be in product management.
• Kofander I am a startup.
• I am thinking of starting a startup in the future.
• I work in a startup company and have no direct role in product planning (for example, I am a back-end developer), but I am curious to know the secret to building a good product?

Where did the story begin?

In the classic definition of network effect, it means that with the addition of each new user, the product/service/user experience becomes more valuable and better for all users [1], [5]. In the following, we will get acquainted with the components and different types of network effects in detail; But to begin with a bit of history of this concept, it is not without merit. The story goes back to 1908: At that time, Theodore Will, Chairman of the Board of AT&T, the largest telephone company of his time, mentioned this concept for the first time in his annual report to the Board of Directors. In this speech, Mr. Will said:

“Without a connection across the line, a telephone is nothing but a toy or a scientific device. This [phone, in this case] is one of the most useless things in the world. Its value depends on the number of connections with other phones, and as the number of connections increases, [its value] also increases.”

Seventy-two years after Mr. Will introduced the concept to the world in this lecture, Mr. Robert Metcalfe, the inventor, and father of the Ethernet protocol, is trying to develop a law for modeling the value of a network that gives It is called Metcalfe’s law.

According to this law, the network value is related to a quadratic number of people in the N² network.

Several years later, in 2017, an MIT researcher named David Reed showed that, in fact, within each network, several network nodes are more closely connected. For example, entrance to computer school in 1996 has a much greater connection with their contemporaries than the university students.

In other words, he believes in each group, many subgroups are deeply related to each other & the value of the network comes from the communication of these members. In fact, the value of the network is explicitly associated with the number of people in these subgroups, known as Reed’s Law.

Recently, in his book, Andrew Chen, former Uber director of growth, tried to provide better modeling of the value of the network, inspired by the growth process of animals that live collectively in nature. For example, the law of mirkat-mirkat is An animal that lives in groups, and Timon said in the cartoon The Lion King — a meerkat (law). According to this modeling, the value of the network increases based on the number of network nodes. Still, from one point onwards, the market reaches a saturation point, and increasing the number of users does not add new value but sometimes has the opposite effect [ Here]. Imagine Snap increasing its number of drivers to understand why this modeling works. This increase in the number of drivers causes the passenger waiting time for a request to decrease from 30 minutes to 5 minutes. Still, from one place to another (for example, the same 5 minutes), the number of drivers increases, and this time decreases significantly. Not for the passenger. Or, for instance, in a social network like Instagram, the first 10 or 20 friends that a user adds have a significant impact on the user experience, but for example, adding a 100m friend is not as important as a 90m friend.

Another reason this modeling is more in line with reality is that as the network grows if there are no proper mechanisms, the possibility of noise or spam increases. In practice, the user experience of a network node may be worse than before. (We will take a closer look at this issue known as the Network Negative Effect)

I mentioned that these laws are not the law and revelation of the house and were merely individuals’ attempts to better model and understand the subject. The practical function of these models is for us to know:

• The narrower and more multifaceted the communication in a network, the greater the power of the network effect. For example, the network effect in a telegram group that communicates with each other is much greater than the power of the net impact of an Instagram influencer who only produces content and uses followers.
• The product’s value does not always grow exponentially with the number of users. The network architecture and the type of communication of the network components are decisive parameters. (We will examine this issue in detail below)

Of course, if you are interested in going deeper into the theory behind these rules, you will find a general report and related links in this article.

Why is it important to understand the mechanism of the effect network?

You who are involved in the development of a digital product have probably heard similar advice at least once:

“Do not worry about building a huge and defensible business against future competitors. Instead, focus now on solving a customer problem. The rest will be solved later.”

It can be said that this sentence is correct in one dimension, And most likely, if you find a severe problem and offer a suitable solution to it, it will probably be possible to find a revenue stream and turn it into a business. But the fact is that to build the next Unicorn and a multibillion-dollar industry, you need more than just solving a problem. The common denominator of multibillion-dollar companies is that they grow exponentially, and in the long run, their competitors can not easily defeat them. That’s why it’s essential to work on an idea from the beginning and put energy into it, at least in theory, that has the potential to grow exponentially and has a strong defense against potential future competitors.

A review of successful startups shows that their most important weapon against competitors is the effect network [ Here].
Also, the results of a study on 336 companies started between 1994 and 2017. They reached a value of more than $ 1 billion: about 35% of these companies’ Network Effect was a significant component of their business model. But the exciting thing is that this 35%, on top of each other together, had about 70% of total company values.
In other words, there is one thing in common between all the major giants that have made the most significant contribution to value production in the world in the last 27 years: Network Effect.

Maybe you have heard it in different languages ​​before and tell yourself that now everyone knows that network effects are significant, and we happen to be working on the method. But it is interesting that despite all the talk and fuss in the startup world about network effects. For example, a report published by one of the world’s leading investment funds, nfx, shows that only 20% of the ideas for investing in the network effect are a significant part of their business.

So it seems that many people still do not understand this concept exactly and are confused with other concepts such as viral growth, economies of scale, branding, or even platform and marketplace. They take. For this reason, I think it is vital for every entrepreneur and product manager to understand this concept precisely. And to explain the product strategy at different stages of the product life cycle, a product manager needs to identify and strengthen the mechanisms and types of network effects in his product. And also have appropriate metrics to measure the health of the network effect in his business.

So what is the difference between what you said and what?

As we said before, when the number of users increases, the product becomes more valuable for all users, we say that this product has a network effect. It is good to know that one of the most important manifestations of this more valuable product is increasing the user engagement rate of your product (Engagement Rate) and increasing the user retention rate (Retention Rate). Demand Side Economies of Scale).

What is the difference between the classic scale advantage and the Supply Side Economies of Scale? Scale advantage [9] means that when a business has many customers, the final cost of producing the brush decreases for various reasons, such as increasing bargaining power with suppliers or controlling the distribution chain. As a result, it has a higher profit margin and a more competitive price for the consumer than its competitors. And well, by now, you have probably realized that this has nothing to do with increasing user usage and making the product more valuable or the same network effect and that the two things are entirely separate.

On the other hand, one thing that is often confused with network effects is the viral effect or viral growth. When an existing user causes a new user to be added to the product for free, we say that the product has a viral boost. But as we said before, Network Effect is about increasing product value for users and increasing the power of defense against competitors as the user community grows. And these two concepts may or may not be completely independent of a product. To make it more transparent, let me give two examples:

• A video trending on virtual networks has a so-called viral growth, but many people who watch that video never go to the page that produced that video. Other examples could be the trend of a game like Angry Birds at a particular time, which gets its initial fever and is no longer so popular after a while. (In this case, we have viral growth, but there is no news of network effects)
• In an online taxi application like Uber, as the number of drivers in a city increases, the average acceptance time of passengers’ requests decreases, and, of course, the user experience of passengers improves. But as we have probably all seen, to attract drivers to a new city, Uber needs to spend a lot of money on advertising and field forces and the reward of introducing a new driver. (In this case, we see that there is a network effect, but we do not have viral growth)

Network definition

Network effect or network effect consists of two words: network and effect.

First, we need to get acquainted with the first part of this word, network (from now on referred to as network), its components, different architectures and features of each, and understand their impact on our product strategy.

Nodes and Links

At the most basic level of a network, it consists of many nodes that are interconnected by links. Nodes are actually participants in the network or the same users, manufacturers, consumers, or even tools such as telephones. In fact, not all nodes in the network have the same value, and some nodes, such as cyber influencers, have a more significant impact on the web. In other words, the more connections and connections a node has, the more valuable that node is in the network. Similarly, links between network members are not of equal value, and the importance of a link is related to parameters such as proximity, stability, and activity of that link. For example, the connection between you and your best friend on a messenger like WhatsApp is far more valuable than the link between you and your mother’s cousin, who has a count on your phone and you never talk to her.

Network density

Network density is the ratio of links to the number of nodes. In general, the higher the network’s density, the higher the value of the network. In fact, the density of the network is not the same everywhere. The practical and usable point for you is to pay attention to the denser parts of your production network. There are more connections and examine the activities and relationships formed in this part. As the product develops, other network nodes will include similar behavior. Generalize this.

Directionality

In some networks, the connections between nodes are two-way. For example, in a messaging platform like WhatsApp, the message is transmitted between two parties. But in some networks, such as Snap, the transfer of money from the passenger to the driver is one-way. Paying attention to the direction of these communications and the different communications types helps prioritize product capabilities.

Clusters

As I mentioned before, network density is not the same everywhere. In fact, most networks are made up of many closely related clusters, and the more prominent these clusters and the connections between them, the greater the value of the network. For example, in a corporate messaging platform like lemon messenger, each company can be a cluster for itself. Different teams in a company have other collections that have more intra-team communication. The product manager should try to strengthen the connection between sets. And allow more prominent groups to form in the network, which increases the value of the network. For example, by adding features such as searching for all organization members or searching by a specific department, it is possible to increase the communication rate between two different teams in an enterprise messenger and have a larger cluster.

Another significant point is that the network is broken into many groups in many products. So-called atomic networks are the leading network and have their own mechanism (for example, different cities for SNAP).
In this case, product manager must separately monitor the health of each cluster and decide on development of each set according to the conditions of that cluster.

Critical mass

Every network has a critical point for it to grow. Then, the value of the grid for the user exceeds the value of the product technology itself compared to the competing product, and the so-called atomic network is built. Before reaching this point, the network is not stable, and in classical literature, it is said that there is an anti-network effect.

Product managers should pay special attention to building these nuclear networks. For example, for a video calling platform such as Zoom, it is enough to form an atomic network of two people who want to communicate with each other. But it may be critical for a social network like Facebook to follow 10 friends. In this case, features such as syncing contact lists must be implemented so that a new user can easily find friends and acquaintances and become a member of an atomic network.

Asymmetry network

In some networks, there are two or more different types of nodes. For example, we have two types of users in the Digital network: buyers and sellers. Or, for instance, in the SNAP network, there are two types of users, driver, and passenger. One part is more difficult to absorb in most multidimensional networks, called the stiff side. Despite that, it is easier to absorb the additional amount. For example, attracting a driver in SNAP is far more than the cost of attracting passengers and the hard side for SNAP drivers. The product manager must be well aware of these differences in the network and usually, prioritize first to facilitate the user experience and help attract more difficult network funds.

Negative Network Effect

In some cases, making the network bigger may cause it to be more valuable and reduce its value. This condition is called a negative network effect. This negative effect can be due to increased use of network users (strengthening links) or increasing the number of nodes. For example, on a city road network, if the number of cars increases by a certain amount, traffic is generated, which makes the network users’ experience worse than before. In the case of digital products, the significant negative effect of networks is the increase in the number of nodes. As a result, communication becomes more complex, so-called network noise or pollution. For example, suppose you follow many people on the LinkedIn network who are not related to you. In that case, Your user feed will be filled with content unrelated to your interests, which will ruin your user experience. Or for example, if you are looking to buy a car in the wall application and a large number of car ads are displayed that are not related to your needs, You will have a worse experience finding the car you want. In this situation, product managers must prevent this from happening by developing various features such as different feed or search algorithms, spam detection algorithms, or personalized suggestions.

Different types of network effects

So far, we are familiar with the components of a network. The different architectures and features of the network and their impact on how the product is developed. Now we want to go to the second part of this word and know the different network effects that may exist in a product. It is important to note that each effect has another power and, of course, another value. There may be several different types of network effects in a product. In this article, other network effects are introduced in more detail. Unfortunately, it does not have that much practical function, and I am bored with this foreign writing. I have tried to review a more practical summary in this article but refer to this article if you are interested in more details.

Direct Network Effect

This type is the most tangible and powerful type of effect network. In this type of increase, the use of the product directly increases the value for users. The direct network effect is divided into intra-aspect (same side) and inter-aspect (cross-side). In the first type, increasing the use of one side of network users directly increases the product’s value for the same side. For example, in a messenger or a social network, the more friends and acquaintances a person has and the more active they are in the network, naturally the more the person uses the network, and the product has more value for that user. In the intermediate type, increasing the usage and number of users on one side raises matters for users on the other side. For example, the more wall advertisers there are, the more valuable the network becomes to those who seek to meet their needs through the wall.

Indirect Network Effect

To explain this kind of network effect, I use physical markets, the so-called stock exchanges or orders of a particular commodity. Consider, for example, a call where everyone is a goldsmith. At first glance, it may seem that increasing the number of gold sellers is not in the interest of other gold sellers because it leads to competition between them. But the more the number and variety of gold shops in a market, the more buyers are attracted to that market, and the more customers favor each gold shop. There is a similar mechanism in digital products, which we refer to as the indirect network effect.

Data Network Effect

In some products, there is no direct impact of user presence, which increases the value of the network and improves the user experience of other users. Instead, algorithms that use user-generated data make the product valuable. For example, in the map and signage router, the more people who are driving with the help of the sign, the more accurate the output of the traffic detection algorithms is, and the more precise the routing of all users. More or less, all digital products use this type of network effect, but it should be noted that the closer this effect is to the product’s primary function, the more critical it is for product managers to pay attention to it.

For example, viewing video content such as the primary function apparatus and network in a site is the main effect between video producers. And consumers and the proposed algorithms are not the central axis of this network but the facilitator’s role in forming the primary connection and preventing the effect of hostile networks. Have.

Technology Performance Network Effect

In some products, the larger the network size, the better the network infrastructure technology, which improves the user experience. For example, in a file transfer system such as Torrent, increasing the number of users makes file transfer faster and more accessible. Or in the bitcoin network, increasing the number of miners increases the processing speed of transactions.

We have reviewed the main points that needed to be said about the network effect. In this article, I tried not to translate the keywords into Persian or to be sure to bring their original English phrases next to them so that you can effortlessly search the internet with these keywords and read more details if you are interested. I hope reading this post in the future can help you better organize your mind and prioritize the development of your products.