One of the ideas of the late nineties’ e-commerce boom was that of disintermediation: that brokers, middlemen, and other layers between buyers and sellers would disappear, squeezed out by price competition. While later observers tried to debunk the idea as a myth, it has never completely gone away. It is easy to see why people are arguing both sides: in some sectors disintermediation is alive and well, while in others it is nothing but a myth. In this note we point out that new understanding of network effects is to a very large extent able to predict where disintermediation will and will not occur.
The concept of disintermediation is a simple one. As the internet increases access to information it reduces inefficiencies in the economy. Brokers exploit these inefficiencies and are therefore a threatened species.
A classic example is flight bookings. Traditionally, airlines’ pricing structures were arcane and not accessible for the untrained. The travel agent provided a useful service, trawling through timetables and complex fare structures to find you a suitable ticket.
However, as the airlines started to sell tickets online, providing simple access to timetables and fares, the value provided by the agent largely disappeared and today most flights are booked online by the passenger.
On the other hand, while large book publishers like Random House allow you to buy their books online through their web sites, those sites are singularly unsuccessful compared with Amazon.com, a middleman in the traditional sense.
Why is this?
We have explored network effects in our research paper Business Platforms: Profiting from Networks. They key point from that discussion is that, for several classes of networks which we describe, the value of bringing together to previously disconnected networks is much, much greater than the simple sum of the value of the networks. The value of the combined network often depends on the square or even the exponential of the number of actors (e.g. buyers and sellers).
Let us look at the flight broker again and consider the networks that he connects. There is certainly a large number of buyers: that is a large network. However, if you want to travel between two cities at roughly a given time, then there are typically only half a dozen airlines from which to choose–and sometimes much fewer.
Compare with the situation of buying books. The network of buyers is even greater but the crucial difference is in the number of sellers: there are many more publishers than airlines. The value of the intermediary is therefore much greater.
That is a lot of the reason why your online flight broker is struggling while Amazon is successful. EBay, the online auction site, is successfully bringing together private buyers with private and commercial sellers. These are huge networks, and EBay’s revenues dwarfs Amazon’s.
As we discussed in the research paper, Amazon.com goes further than simply connecting the networks. By allowing customers to provide content and by providing grouping and community features both on its own site and externally, Amazon is starting to move its network from Metcalfe’s law to Reed’s law.
To quickly recap the discussion from the research paper: Metcalfe’s law is good for networks that allows paired connections and says that the value of the network grows with the square of the number of actors. Reed’s law holds for networks that allow group formation and states that the value of the network grows proportional to the exponential of the number of actors (i.e. much faster that Metcalfe’s law).
Communities are key to delivering massive growth in the value of the network and to delivering market dominance by providing the platform that connects networks.
That is why Amazon’s moves are so important, and this is examined at greater depth in the research paper.
That also explains what travel agents must do to re-intermediate, to suggest a word for this business trend, their business and win back their customers.
First, and this is already happening, they must bring together larger networks by catering to the whole travel experience. There may be only six airlines to choose from, but add the hotels, the car hires, the excursions, the travel to and from the airport, the insurance, and all the other services and products that are required for a vacation or business trip, and you have much greater complexity and therefore much higher potential value for the broker.
Second, use the collaborative and community effects to improve the class of network that you are connecting. I see travel guides and travel agents merging and becoming a single resource for the traveler. By providing additional information and by allowing the travelers to exchange knowledge, tips, ideas, and experiences, the network moves toward Reed’s law.
That, surely, is the future. It will not stop airlines from selling tickets directly, but it will re-establish the travel agent as a platform for travel experiences and clarify the value that he brings.
One word of caution. In the research paper we argue that because the value of combining the networks grow so fast, platforms are natural monopolies. You may see one platform for business travel and one for leisure (or perhaps a few: high-end travel versus budget experiences, for example), and you may see a limited geographical competition. But the final number of successful companies in this business will be very small.
See Also
- Our research paper: Business Platforms: Profiting from Networks