There's More Than One Way To Do It

Designing and Operating a Successful IT Company Evaluation System, Learning from Fujitsu's Collapse Case Study

Notes from reading "Fujitsu from the inside".

I read "Fujitsu from the inside"

This is an old book from 2004, but I read "The Collapse of Fujitsu's 'Performance-Based Management' from the Inside" , which describes the inside story of Fujitsu's performance-based management system.

The book describes how the performance-based system was introduced at Fujitsu, which originally had a seniority system, and what went wrong. Although it is only from one author's point of view, as you can see from the title, the tone of the book is that most of the things did not work out.

Fujitsu Group is a large company with 124,000 regular employees and 12,000 non-regular employees in FY2021. While there are many articles on the organizational structure and the difficulties that startups face when they grow rapidly, I felt that it was rare to find a detailed case study of a large company.

The table of contents is as follows

  1. The business performance that took a nosedive
  2. How the employees lost their motivation
  3. Total irresponsibility within the company
  4. "Performance-based system" and corporate culture
  5. The dark side of HR
  6. Establishing a Japanese-style "performance-based" corporate culture

After listing a number of problems, the author concludes with his own suggestions. I agree with some of the author's suggestions and disagree with others, so after briefly summarizing the problems that arose, I would like to offer my own suggestions, ignoring the author's suggestions.

Note that this is an old book, about 20 years old, and it may well be improved now. In 2020, Fujitsu has announced a personnel system that will pay 35 million yen a year to advanced IT professional.

Gradual introduction of evaluation system

In 1993, Fujitsu introduced a performance-based target management system. The entire system was not changed at once, but it was introduced to only managers.

Since the idea of performance-based management includes wage cuts, the introduction of the system from the management level may mean that higher wages = higher effectiveness. Discretionary labor systems, grading systems, etc. are similarly introduced from the upper levels.

Fujitsu's evaluation system is based on grades and target management. The grades are roughly as follows.

Grade 3: High school graduate
Grade 4: University graduate
Grade 5: Graduate school graduate
Grade 6 Chief, section chief
Grade 7 Section manager
Grade 8 Department manager
Grade 9 Division manager
Grade 10 General Manager

The goal management is based on a five-point scale of SA, A, B, C, and E. E is rarely given, according to the company.

Problems that have occurred

There were a surprisingly large number of problems, but some of the most common problems are

  • Inappropriate goal setting
  • Evaluation is unfair
  • Unestablished evaluation methods for managers
  • Relative evaluation

Inappropriate goal setting

It seems that although goals were set, but they were rarely checked within the department. It was evaluated by personnel. Since it was a large company, HR could not read the goals of all members. As a result, the evaluation was based on the sense of effort by looking at overtime, sick leave, etc. without reading the text.

Evaluation is unfair

It seems that the cooperation between the department manager and the section manager is weak, and the section manager, who knows the work of the members best, was not allowed to attend the evaluation meetings. There seemed to be an advantage in evaluation if you belonged to personnel. There were variations between departments.

Unestablished evaluation method for managers

There was no demotion system, and due to the change from a seniority-based system to a performance-based system, those who were promoted sufficiently through seniority were able to stay after the evaluation system was changed. There was a situation where younger employees were not be able to be promoted. In addition, the evaluation method for managers themselves was not yet established.

Relative evaluation

In target management, the ratio of SA, A, B, and C was predetermined. As a result, only subordinates of senior managers were given preferential treatment, and political factors for decision making increased.

Later, the evaluation system was formally changed to absolute value evaluation, but it was meaningless because the budget for bonus payment was fixed.

How can we prevent problems from happening?

Reduce the weight of goal evaluation

Although some companies have an annual salary system, traditionally, the HR system in Japanese companies separates base salary and bonuses. Base salary is often based on grade and bonuses are based on goal setting.

For Toyota in 2021, the bonus is 6.0 months, which is 1/3 of the annual salary. In Fujitsu's case, it seems that goal setting had a relatively large impact on annual income.

I thought it would be good to have a smaller impact of goal setting on annual income if the company was transitioning from a seniority system or was newly established from a place where there was no evaluation system.

The evaluation is based on a person's basic skills (competencies) and it is adjusted by performance. An annual salary system without bonuses is basic skills oriented. If it was completely determined by the number of contracts signed, it is incentive oriented. Since seniority is based on a person's potential, I think it would be less likely to cause problems if the ratio is shifted gradually rather than all at once to a set target.

In the case of startups, it is possible to pivot after six months, and the effectiveness of goal setting is often questionable. If the company is a 0=>1 startup, it is likely that the goals themselves will change and not be operational. In that case, I think it is fine to set goals as a tool to spread the vision, etc., and reflect them a little in actual compensation.

Some companies make a resume once a year, give it to a recruitment agent, and then decide on an evaluation based on the annual salary of the person when he/she changes jobs.

Although there are fads and fashions in technology, unlike sales and marketing, I don't think the contribution of an engineer to sales is not too affected by a single year's effort. I think it is a mistake to be too blind to set goals when using a performance-based system. You should also consider a professional evaluation of the person's potential.

The section manager's manager should sample regularly

In that sense, I think it would be good to keep the members' day-to-day work in a state where their manager's manager can properly say what the members' day-to-day work is. It is important for the evaluation committee to evaluate goal setting neutrally, but I wonder if it is difficult to be fair if the evaluation is rushed only at the evaluation time. In particular, if the competency part of the evaluation is left to the manager's discretion, it will lead to variations among divisions.

For example, if a department consists of five sections with eight members per team, 40 members as total, the department manager can get a feel for the atmosphere by attending the morning meetings of each section from Monday to Friday. If there is a low-performing team, he or she may be able to improve the way things are done. It is also a good idea to have 1-on-1s regularly, skipping over the hierarchy.

Consider also the evaluation of managers by their members

Evaluating managers is a more difficult issue than to the members. If a department is in charge of numbers, it is possible to use the overall total as the evaluation, but this is not always the case.

Some say that 360-degree evaluations between members are a bit too casual and don't make much difference, but I think it is better to take evaluations from members to managers. Recently, due to the labor shortage, some companies reflect the evaluation of the workability of their members in the remuneration of their executives.

It is surprising how many teams have a dark side, even if they seem to be doing well from the outside. Sometimes the better managers in term of sales achievement are more harsh with their members.

Prevent the explosion of evaluations by Two-point evaluations

Even though we try to ensure fairness in evaluations through the evaluation committee, I think we are creatures who try to set loose goals. In that sense, including commitment and target in goal setting may be a good method.

The goal is divided into two goals in the manner of a two-point estimate.

Commit is a goal that can be reached if you do your best, and if a perfect score is 100 points, you will get 50 points if you achieve your commit.

The target is a goal that you can achieve if you work really well. You can get 100 points if you achieve target. For instance, it is about a 50% increase in commitments.

Changing from a one-point estimate to a two-point estimate has the effect of slightly reducing the blurriness of the evaluation setting.

If you are still concerned about the budget in absolute value evaluation, I think it is better to move to a competency-oriented evaluation system as mentioned above.


This is an old book, but it was worth reading. It is an interesting case study, so please read it.