How to stop worrying about inefficient employees

Epigraph. “Criticism is not accepted at all, only alternative construction is accepted. If it is better than mine, then I will take it as more correct. ”

The article was written under the impression of the article Motivation of Karate: How to stop worrying about inefficient employees? .

The article provides an alternative view on how to build a bonus system. Partially, the technique was tested on its own skin and showed its high efficiency in the applied area (not software development).
')

^{picture to attract attention}

Disclaimer: No matter how the bonuses are charged, there will always be those who will think that this is unfair. I am simply suggesting another way to do this.

Each task has a cost for the business (as an expense) and has a benefit for the business (as an estimate of the profit the company will receive if the task is completed). Usually, tasks that increase profits are non-standard, since if there were standard, then everyone would do standard tasks and everyone would be fine with the IT business. Empirically, this is not the case, many startups and even mature companies just fail.

But even for typical tasks, such an award system will work poorly. A few thoughts from the adjacent area. I went to school and received money from my parents for grades. Homework is a typical subject of a spherical task in terms of man hours. You take the condition, apply the rules, get the result. (An important historical note. The system does not work at all at the institute)

Let's compare the techniques for me and the one that was in the article.

1. There is a specific lesson for which the task is given. He will be checked and evaluated. For 5 I will get a lot. For 4 little or nothing. For 3 and 2 will be fine. What is my motivation? There is a motivation to do the best. There is a motivation to do the fastest, because the rest of the time I can play on my computer.
   
   And what is the motivation for your system? First, because The system introduced a deliberately wrong lever, I will try my best to increase the estimated deadlines for implementation. Secondly, there is an implicit opportunity to hang from the task. She is always at work and at school. Indicate to the authorities that there is a higher priority task that you specifically invented. Do it easier and faster, and you can not do some garbage. Even worse! I can add a bug to the repository. Start it in the tracker, set a time for it, make it and get points.
   
   
2. There are priorities between subjects. For technical subjects give more, for humanitarian less. Business (in this case, parents) very clearly set priorities in monetary terms. Therefore, I will try more on technical subjects - there is always a chance that if I missed something, then they will ask me. Therefore, I will spend more time re-reading a book on mathematics, and I will not read War and Peace. If I do not get 5 in math, but I get 4, then the loss of profit will be large, and if in the literature, the situation is more indifferent. Yes, I will strive to get 5, but if I get 4, I will not be very upset.
   
   In the methodology of the article, the connection with the business is, to put it mildly, strange. There are specific tasks that will increase the company's profits. Some more, some less. You have a team. No matter who makes the task - the main thing is that it is done and the profit has grown. Your monthly expenses for a team are approximately fixed, as opposed to profit. How to fix this, see p.3.
   
   
3. There is a final score in the report card for the period. I can endlessly cheat with the estimates in the process, but if I fill up a couple of important controls, I will lose ~ 50% of the profits because somewhere half the money I get for the current estimates, and half just for one tsiferku in the report card. Therefore, I am very motivated to get the final result, but not an intermediate one. Needless to say that you simply do not have this part. Awards at the end of the year are not about that at all.

Adaptation techniques for the team is not so trivial, but try to figure something out.

First, we will collect information about what we have, what we want to build and what to avoid:

1. It would be nice if the developers instead of “oooh, one more task” or “how good it is that I have no tasks, I don’t do anything, and the money is dripping,” they said, “here’s the result, come on,” “come on make the product better ”,“ and I used our website and found such a bug ”(reservation: it’s very difficult for staff to deliberately introduce bugs, if it’s systemic, but it’s very easy if it’s marginal).
   
   
2. We have a specific lead, which is responsible for the direction. He has some information about what a business needs and there is a team as a means of achieving business goals.
   
   
3. Does it matter how many people do the assignment? If 2 tasks are done by 2 people, then it takes 1 hour. If at first 2 people do one task together and then another, then it will be 1 hour too. But there is some difference. The more people will do the task (up to a certain limit), the more inaccuracies can be noticed.
   
   
4. Nobody loves when they are underpassed by the RFP, but everyone loves when they are paid extra.
   
   
5. Each system may be perfect on paper, but there may be no way to build it.

Now we try to collect.

Note. Here I will mix lead and PM to one role - “chief”.

1. There is a specific period. Let the quarter. There is a quarterly earnings report. This is an assessment of the work of the entire company and the IT team in particular. For an increase in the company's profits, a bonus is due. All involved. But to share the bonus can be proportional to how each participant invested in the result. OK. Business gives 10% of the increase in revenue per quarter as a bonus.
   
   
2. Let us have one lead. How much should get from this lead - a separate complex issue, you need to think. But somewhere 90% of the money should get the team that did everything. The lead directly influences how he distributes tasks and, as a result, how many each will receive. Its tasks are: distribute tasks relatively evenly so that more can be done, distribute tasks so that they are done as efficiently as possible, to occupy idlers (in the good sense of the word, the fact is that different employees idle at different stages of the company work, this is normal and not correctly deal with this, but to eliminate the narrow link (see Goldratt - Goal 1)).
   
   
3. And here we need metrics to share this bonus. During this period of time there were tasks that needed to be done. There were people who participated in these tasks. It is logical that the more a person participated - the more he should receive. But here IMHO you need to make one small detail that no one else does. Add something like a sigmoidal function for a part that corresponds to too much reward. Why is it important. First, the problem is to determine how much the lead gets. Here it becomes clear that he simply will not receive more than a certain amount according to the formula, because it makes sense to pull the blanket over yourself all the less, the more you pull it over yourself, and not proportionally. Secondly, there is risk management, if the team has one knight on a white horse, then bus factor will tend to 1, and this should be avoided. Understanding asymptotes improves the atmosphere. You are not trying to climb out of the skin and jump over your head to get another reward, respectively, you burn less. In the end, it's better for the employee and the lida.
   
   
   
   
   Example of salary in $, as a function of the work done.
   
   
4. In addition to the bonus from increasing profits, you need to add a bonus from current profits. This is important because if the company is stable, the system simply will not work. No examples of positive reinforcement.
   
   
5. Some fixed part of the bonus is vital to make subjective. This part should reflect the attitude of the manager to how this particular employee works.
   
   
6. New employee will be quite difficult to hire. If you offer him a rate below the market + bonus, then everyone will say “we know, floated”, and if you give 4k $ + bonus immediately, then the authorities will already doubt if “we should take a more effective manager”. Therefore, there should be a smooth integration of a new employee into the company's culture. The first 3 months - conditional probation. There is only a fixed rate. For the next 3 months, bonuses will be distributed to the employee as well as to all. But stripped down 2 times. Here the employee can feel the power of motivation bonuses. After the first receipt of bonuses, the employee is announced that in general he can now switch from a large fixed rate and a small bonus spread to a lower fixed rate and a larger bonus spread. There are several options discussed with the business.
   
   
7. Sly gouging will not work. If you constantly do the minimum "just to get a fixed rate", then it will not last long. At the moment when they offer to switch to safe fixed rate, the cunning dolt will remain on the old terms, but this automatically attracts the attention of the immediate superior. Here is such a feedback.
   
   
8. By construction, personal inefficiency is suppressed weakly. Those. if you do less than everything - ok, at this stage you are a slacker, but then you may have the opportunity to prove yourself on another task. If you can not express yourself in any way - goodbye. If personal inefficiency develops into something more and already prevents parts of the team from working normally (which affects their bonuses), then “goodbye” may come much earlier.
   
   
9. A subordinate may well receive more than his immediate superior. This is normal, moreover, it allows for career growth without changing the office or moving to the level of incompetence (see Peter Principle )

A few words in the end. Unlike the author of the original article, this technique was not used in its entirety anywhere in software development. This article can be viewed as a detailed commentary on the original article.