Peer-to-peer:

Assemble four workers into a peer-to-peer topology. Each team member acts as an independent participant in a perfect free market.

There are twelve information channels.

High redundancy: workers send duplicate data to each other.

High contention: duplicate data can be misinterpreted or out of date.

High latency: most data travels three hops.

Peer-to-peer systems do not scale well because transaction costs rise exponentially as new nodes are added.

Hierarchical:

Now assemble the same four workers into an organization (a corporation) and add a manager who partitions information and redirects it to the proper destination. Team size increases by 25%, but information channels are reduced 33%.

Now there are eight information channels.

Low redundancy: each worker has only the necessary data for their job.

Low contention: The manager is the sole source of project information.

Lower latency: information can travel a maximum of two hops.

The hierarchical organization substantially reduces errors, conflicts, unnecessary duplication and decreases reaction time by reducing transaction costs.

Click to invert this theory