SOME BACKGROUND ON PEER-TO-PEER NETWORKS

At the heart of a peer-to-peer (P2P) network architecture lies the idea that entities or peers can connect directly with each other. This is trivial in a sense, as one might argue that all interactions between people are direct in nature.

However, in a large set of peers, search costs are high. If many sellers and buyers exist in a market, it would take forever to meet every single person with whom they might want to enter into a transaction. In this case, a central authority may take on the burden of pooling all peers on one side of the transaction—such as the sell side—and represent itself as one peer that the other side can buy from. An example would be a supermarket; all producers sell directly to the supermarket, which becomes the convenient place to go for all buyers of produce. However, this convenience comes with several disadvantages. For instance, the central authority has a lot of power over buyers, as it is now the monopolist in the market with the power to set prices. If a supermarket has managed to lobby for laws that prevent individual sellers from selling directly to buyers, both sides are at the central authority’s mercy. Sellers have only one party to sell to, and buyers have only one party to buy from (Figure 2.6). This form of interaction is no longer a peer-to-peer network.

Even though the central authority often argues that the convenience he provides to cus­tomers comes at a high cost which justifies a high price, a centralized system is rife with the

FIGURE 2.5 Peer-to-peer networking with direct and indirect connections

FIGURE 2.6 Centralized systems architecture, a supplier-oriented marketplace

potential for abuse.

For direct connection to be economical, peers must have information about each other. When all peers already know each other, this is simple, but the burden of finding the right peer quickly increases with the number of peers in a network. At the same time, a system with many peers benefits strongly from network effects: the more individual peers exist, the more permutations for connections exist, which gives buyers more variety and sellers a bigger market. However, there must be a mechanism or platform for peers to gather information about each other, which will facilitate connections. As a trade-off, peer-to-peer platforms allow only bilateral connections among lenders and borrowers, not linkages via other peers. As you can see in Figure 2.7, peer 6 can have only bilateral links with all others in the system (peers 1, 2, 3, 4, 5 and 7).

The scale of centralized structures that we saw in Figure 2.6 had clear advantages for users who had no other way to access a market or gain information about potential trading partners. However, with the emergence of the biggest peer-to-peer network in history, the internet, it has become easier to collect, structure and make available searchable information about a large number of peers. In a sense, online peer-to-peer networks reintroduce the ancient way of direct interaction that humans used in trade hundreds of years ago.

Whenever we bring up the term peer-to-peer, most people first think of illegal file sharing services, such as Napster. In technical terms, Napster was a first-generation peer-to-peer system: it still relied on a centralized server structure that eventually proved its Achilles heel. The system could have detected illegal file sharing activity on its infrastructure but chose to ignore it. Subsequent protocols, such as Gnutella or BitTorrent, did away with a central directory.

FIGURE 2.7 Peer-to-peer platforms allow only bilateral linkage among peers

uses that also worked as a commercial business model. It expanded into other arenas, such as television, with P2PTV.¹²

2.4.1 Disintermediation or re-intermediation?

By encroaching on the turf of the established financial sector, FinTech startups profess to disintermediate the established players. Some platforms connect lenders and borrowers directly in a marketplace, so they make the gatekeeper obsolete, they say. Providers of mobile wallets and payment processors profess to abolish credit card companies. Cryptocurrencies aim to unseat central banks. However, even though a platform may be neutral, it will reroute traffic from elsewhere through its proprietary infrastructure. It is therefore more accurate to speak of a re-intermediation, not a disintermediation. Even though some of the startups may disagree, if their business models prove successful, there will simply be a new sheriff in town, in the form of the platform, ruling over a particular niche. Over time, these new incumbents will suffer a similar fate like the established players they unseated. They will have to be vigilant to stay agile, continuing to innovate.

The same has happened in other sectors, where technology has upended established companies. The book trade, for example, has seen widespread disruption, where traditional retailers of books and their publishers gave up a large market share to online marketplaces.

2.4.2 Infomediaries, intermediary-oriented marketplaces, and the information value chain

Information is a powerful enabler of peer-to-peer networks, and the spread of online transac­tions has given power to a new kind of intermediary: the infomediary. Authors John Hagel and Jeffrey Rayport coined the term to describe the function of a central negotiator who col­lects information about customers and helps them identify the best deals. Negotiations with consumers for information cost time and money, and the information is often messy and hard to compare. At the same time, many companies may collect information about customers, but they organize them poorly. When Hagel and Rayport wrote about infomediaries in 1997, they noted the process of collecting consumer information was already under way. However, they pointed out that it might take several years to play out across broad segments of customers and products.¹³ Advances in online commerce and technology have pushed the necessity for information about counterparties into the mainstream. Consumers are also increasingly willing to share private information, such as their financial data, when they can benefit from it. When we look at the onboarding process of marketplace lending platforms, it becomes immediately obvious that information is at the core of this business.

When we compare the model of the infomediary to the supermarket we just discussed, we see how the supermarket could transform its business model to become an infomediary: instead of purchasing produce from sellers, it could simply collect information about their inventory in a database. When a buyer enters the store—offline or online—she could query that database and see which seller offers what she wants at this particular point in time. The buyer would then pay for the information and collect the products directly at the seller's location. Of course, markets have been infomediaries for quite a while. In parallel to the core business of selling produce, they have always collected information about buyers. Nevertheless, now that shopping increasingly shifts online, information has become more valuable and more important. Buyers search a database of products on a website where they complete the transaction. The seller receives a notification of a sale and the shipping address of the buyer to fulfill the purchase. This insight is by no means revolutionary because many marketplaces that operate this way exist, such as eBay, Amazon, or Alibaba. They are what author Dave Chaffey calls intermediary-oriented marketplaces, contrary to supplier-oriented marketplaces (which we saw in Figure 2.6) and buyer-oriented marketplaces. The organization that finances and runs the platform determines the orientation of a marketplace. In supplier-oriented marketplaces, the sellers of a good or service invest in information technology to attract buyers. In finance, this is what a bank would do by making available loan applications online, or by investing in online banking. When firms invest in setting up buyer-oriented marketplaces, they do this with the goal of reducing their procurement costs.

2.5