Mogic considers

With the combination of a small number of people + software + servers and robots
We are promoting a new era of company management.
We hope to share part of this process with you in this corner.

Representative Director Yoichi Yamane

June 25, 2024

Rewrite the organization as information

It occurred to me to do a thought experiment that is typical of IT companies.

First, assume that everything in the organization can be converted into a "flow of information".

In short, everything is considered "information flow," including the contents of e-mails and meetings, casual daily conversations, teaching and learning, hiring new people, getting to know someone you don't know, preparing quotations and contracts, issuing invoices and checking payments, managing cash flow, setting up office facilities and work rules, checking attendance time, enhancing security, and so on. Checking office equipment and work rules, checking time and attendance, enhancing security, etc., are all considered "information flow.

The question then becomes how best to route (design routes) if one is in a position to have a bird's eye view of all information.

This does not mean that because you are the head of the organization, you are authorized to view all information, but it is a hypothetical situation where you really have access to all information.

In a normal work site, it is enough to process the information coming from one person's point of view sequentially, but in this case, it is difficult to keep track of everyone's information at the same time.

Even if it were for one day in an organization of 10 people, the simple addition of all the information volumes would be enormous.

Furthermore, as time passes, current information creates new information as a chain of new information, which, if left alone, will accumulate linearly (whether in a straight line or curve).

The image is that if nothing is done, a lot of unread mail will accumulate.

Now, what should we do?

First, consider the method of selection & prioritization.

The granularity of information is determined in advance and divided into appropriate sizes.

Prioritize the divisions based on the measure of their impact on the organization as a whole.

A category that affects more than one category is considered higher than one category.

Select from the top of the priority list and discard the lower ranges.

This way, the amount of information can be reduced, since the resource of attention can be devoted only to the top priorities.

However, risks remain.

What may seem to have a small impact from the perspective of the organization as a whole may miss the larger impact from the perspective of society outside the organization.

We will call this the top-down "single viewpoint trap.

Second, consider node & filtering techniques.

Determine the important nodes (nodes) that will serve as hubs and spokes, and establish filtering (screening) functions at these nodes.

In layman's terms, information is gathered from the members to the head of the department, and the head of the department makes a comprehensive judgment and then gives the information to a higher node.

If the node takes a hierarchical structure, the more steps you go through, the more the information decays.

This is also a risk.

It is easy to make partially-optimal decisions only around important nodes, an issue that can easily delay information when there is a hierarchy or when nodes are far from each other.

Partially optimized node-by-node information arrives at different times, and as a result, it takes longer to make a policy decision as a whole, which may magnify the bad effects.

We will call this the bottom-up "multiple node trap.

Third, consider redundancy & compression techniques.

It applies what is used in information theory to minimize the number of packets to be sent.

Compresses data by treating similar portions of all information as identical.

If you have information that X = {A, A', A'', A''', B, B'}, it is shorter to do something like {A*4, B*2}.

Therefore, if all information that is similar to each other is considered redundant and identical, it can be compressed, thus reducing the amount of information while maintaining the overall structure.

This also has risks.

If too many parts are made similar, the compression ratio may be too high and the information may not be understood at the destination.

On the other hand, if the resolution is increased too much, there is no place to compress.

In other words, it is difficult to determine the criteria for what constitutes similarity or not.

We will call this the "redundancy level trap" due to resolution.

For reasons of space limitations, I cannot expand further, so I will force my way into a summary.

Since there is an overwhelming increase in information within organizations compared to 10 or 20 years ago, I propose that we should work on organizational networks as informatics rather than organizational charts as geometry or taxonomy, without being afraid of various traps when new problems arise.