iTranslated by AI
Organizational Dynamics in Technology Selection
TL;DR
- I've illustrated the criteria for technology selection and their outcomes in simple diagrams.
- While no decision is strictly right or wrong, we should at least avoid decisions based on weak grounds with little consensus, such as a "command from the boss."
- Finally, I've listed some examples of selection criteria.
Information Graph
I've experimentally used Mermaid to represent my points in a graph structure. Perhaps keeping this in mind while reading the following content will make it easier to understand.
Introduction
Today, I posted the following tweet.
It received quite a bit of response, so I’d like to explore this further.
How Technology is Selected
In that tweet, I focused on "whether to automate infrastructure," but it could be anything, such as introducing k8s.
I will summarize what is thought about—and how we should think—when a relatively large-scale technology selection is needed in an organization.
I called it "Dynamics" because I imagined the central issue as something like a sphere, with various forces being applied to it from the left.
The right side shows the destination resulting from these forces; moving towards the top-right represents adoption, while moving towards the bottom-right represents rejection.
Furthermore, regarding the forces on the left, the opinions and thoughts of those pushing to adopt the technology (the pro-adoption vector) point towards the top-right, while the views of those who don't want to adopt it point towards the bottom-right.
Additionally, I’ll incorporate a vector-like concept where the size of the arrows represents the magnitude of the force.
When the Pro-Adoption Forces are Strong
First, let's look at a case where the pro-adoption forces are dominant.
I’ve only drawn a few arrows here, but there are various factors to consider.
-
Engineer Interest (⤴)
Based on the stereotype that engineers love new things, this vector assumes they would want to adopt something new and useful that doesn't exist in the current environment.
-
Benefits (⤴)
This vector is based on the premise that the technology in question provides specific benefits, such as reduced construction costs, shortened development time, and other various advantages.
-
Learning Cost (⤵)
This vector represents the negative aspects, such as the need to learn new things requiring additional man-hours, or the necessity of temporarily operating both old and new environments. Conducting workshops costs money, and the automation mechanism itself may incur costs.
Whatever the reasons may be, if the cumulative vector for adoption is stronger, it is considered that the outcome will move towards "Adoption" (the top-right direction).
When the Rejection Forces are Strong
This is similar to the previous case, but where the forces against adoption are dominant.
-
Status quo is fine (⤵)
I've phrased it a bit casually, but this vector represents the perspective in some organizations that "things are fine as they are; we've provided value this way so far, so there's no need to change."
The number of opinions and the magnitude of these vectors will vary by organization, but in any case, if the rejection vector is stronger, the outcome is expected to move toward "Rejection" (the bottom-right direction).
When the Boss's Opinion is Dominant
The previous diagrams showed cases where decisions were made based on "opinions" that contribute to the evaluation criteria.
On the other hand, in organizations where terms like "anticipating the boss's wishes" are common, there may be cases where the opinion of someone in a high-ranking position is dominant.
The label on the arrow simply says "A word from the boss," indicating that the actual content of the statement may be irrelevant to the rational decision-making process.
The fact that this single word is prioritized is represented by a larger arrow.
Of course, the final decision-maker is the person in charge, which is why it is crucial to provide them with the correct input so they can make an informed decision.
What I want to emphasize here is that we should avoid situations where something is adopted or rejected simply "because the boss said so."
Examples of Decision Criteria
While some of these overlap with what I wrote on the arrows above, here are a few examples of factors to consider when making a decision.
-
Engineer Interest
I believe engineer interest is very important. If interest alone is the reason for adoption, it could lead to future issues regarding what to do once people get bored. On the other hand, if a technology is adopted solely because "it makes money," engineers may not find it interesting, which can lead to a failure in implementation. I believe that engineer interest serves as a catalyst for understanding what kind of benefits a technology can bring.
-
Cost Reduction
Cost reduction is also an important factor in technology adoption. To be honest, cost reduction is straightforward, and results appear quickly. Especially for IT systems that aren't directly related to increasing revenue, it is possible that cost reduction is the only thing that can be achieved.
-
Revenue Increase
Similar to cost reduction, this contributes to return on investment, but it is more challenging. There is almost no guarantee that adopting a technology will immediately increase sales; in many cases, only indirect contributions can be measured. In tech companies, adopting technology may lead directly to service improvements and increased revenue, but for non-tech companies in industries like retail or manufacturing, only top-tier companies seem to have established systems where IT adoption directly impacts sales.
-
Learning Cost
Introducing any new technology comes with a learning cost. For example, with k8s, it’s not something that members accustomed to VM-based operations can manage overnight. They need to acquire knowledge through books, workshops, or certifications. Furthermore, simply telling members to learn often fails to motivate them; correct incentives linked to organizational and individual KPIs are necessary. While in Western countries, it may be relatively easy to supplement necessary learning costs by hiring temporary staff, Japan offers less flexibility from an employer's perspective, requiring a strategy on how current members and the organization should grow.
-
Construction and Operation Costs
While not applicable to everything, any automation mechanism will involve corresponding construction costs. It is also easy to predict that maintenance and operation costs will follow. Specifically regarding automation, one must carefully consider whether the benefits outweigh these construction and operation costs. If the infrastructure architecture varies significantly every time, requiring manual adjustments outside the automation framework, the cost of automation is likely to be higher.
-
Conservative Organization
"Status quo is fine" can be described as a conservative organization. This is a mindset that dislikes change and wants to stay as-is. As is often said, if everyone—including all competitors—stays the same, then the "status quo" remains. However, the moment a competitor takes even one step forward, it is no longer the "status quo." You are only maintaining the status quo if you can keep pace with others. Considering that the world is constantly advancing and improving, some form of constant improvement will always be required.
Summary
This post came about because I had some fuzzy thoughts that I tried to put into diagrams; I found it interesting enough to dig a little deeper.
In business decisions, there are no absolute right or wrong answers, and no decision can be completely shielded from criticism.
That said, we want to avoid situations where excessively vague decisions continue, leading to a loss of morale among the members who actually handle the implementation, construction, and operation.
Discussion