Sunday, April 30, 2017

System Behavior Charts

This is a link to a (mostly) technical article, but with huge implications for decision making.

Measuring various metrics is the cornerstone of informed, data-driven (or at least data aware) management. However, acting upon these metrics without a proper understanding of the underlying system can only amplify erratic behaviors or lead to hard to contain side effects.

Experts recommend not doing anything as a better alternative to uninformed decision making, but not doing anything is rarely an option for a company eager to perform. Therefore, in our competitive landscape, now more than ever I believe a good understanding of how to read and interpret data is critical for sustainable continuous improvement.

Here are some guidelines: https://alexandrugris.github.io/statistics/2017/04/27/short-intro-to-reading-data.html

Wednesday, April 26, 2017

Notes On Management

On working with multiple managers

Sometimes it happens that you are in a matrix organization and might have two managers with (somewhat) overlapping areas of responsibility. This is not necessary a bad thing. For once, it is a great opportunity to open your eyes and learn from the situation, either as the employee with two managers or as one of the two managers of the employee. As everyone is bringing different expertise and points of view to the table, higher quality insights, learning and decisions will surface. However, it might be at times painful, especially when trust is not fully formed between the involved parties.

Here are some guidelines I found useful:
  • It is critical for the two managers to get along very well and not to try to get advantages one over the other. Think child with two parents. Even for the most well intended people, misalignment will occur and trust will be eroded if the two managers don't take special care to continuously build and cherish a trust relationship between themselves. Relationships are not to be taken for granted. Work is needed to build a maintain a relationship between two people.
  • It is highly recommended for the two managers to align between themselves on the roles and responsibilities of each other and re-discuss these from time to time. Understanding might drift as new situations and challenges arise.
  • Try to avoid the communication triangle and rather aim for the star. 


In triangle communication, each two people have a separate communication channel, meaning they have separate meetings without a common tripartite alignment process. In the star communication, all three meet regularly all together and they reconfirm their understanding. This latter setup drastically lowers the risk of misunderstanding and / or of the child with two conflicting parents situation.

On what I call the airplane model of organization design


  


I have a strong belief that people have a great capacity to take the right decisions provided that they have the freedom to do so and the right information easily accessible. When I am thinking about organization design I am centering it around what I call "the airplane model". The model itself is simple: the individual is in its center (the pilot). The rest of the organization is there to support him take the right decisions (the instrument panels in the cockpit, easily accessible). Beside the empowerment and motivation that comes from such a setup, of being in control of the course of your work, the model has some other less obvious advantages:
  • It scales. The pilot is everyone and everyone is flying its own plane supported by others. The programmer codes while the team-lead helps him get access to knowledge (architects, websites, classes, conferences), tools (task boards, IDEs, code metrics, continuous integration servers) and goals. The team-lead leads while his manager offers insights, perspectives, a sounding board, access to resources, confidence, the program manager access to configured Jiras, QA to product metrics and so on. 
  • It mandates collaboration and knowledge sharing because the underlying assumption is that nobody flies blind. Decisions are fully owned, but checking the instruments (discuss, check metrics) is a must.
  • It is a quest for continuous improvement: always simplify access to information, always discuss, interpret metrics or discard irrelevant ones. Bad decisions can happen and they are expected as long as we learn from them, but ignorance is not permitted. I personally am OK with any decision as long as I have the certitude that all factors have been properly weighted.

On compound interest and organization design

In any management endeavor, initial conditions matter greatly but they can be beaten given some conditions occur. However, beating the averages is a hard job. Given the fact that growth and decision making at any point in time are random variables normally distributed around the system's average, usually the system beats the individuals, thus organization design is critical for a successful growth.

A little bit about compound interest and exponential growth:

Let's consider the following two cases of growth:

First case: 

Initial condition: 1 and 4 respectively, same exponential growth of 1.7. It is obvious that the initial 1:4 factor is preserved and the difference is only amplified as time passes.

Second case:

Same initial condition 1:4, but this time exponential growth is set to 2 to 1.7. After roughly 10 periods, the initial advantage of the second series is completely lost and from here on the difference grows at an exponential rate in favor of the first. This means that it takes time to recover, but once the gap is recovered, the first series obliterates the second.



But how do we translate these results to day to day management thinking?

Obviously, the initial conditions matters. A company or a team with more funding, better trained people, more connections or that simply starts earlier has a clear advantage over a less fortunate competitor. Because the less fortunate competitor needs time to recover, he may simply run out of business before the gap is closed, even in case of perfect execution. However, once the gap is closed, better execution gives much higher yields.

So the main question is, provided that we have enough funding to survive, how do we increase the exponent? I think the exponent has three defining characteristics: ability to learn, to execute and to incorporate learning back into execution. To learn means to experiment, to play. That means a culture willing to tolerate failure and learn from it and willing to invest in people - to give them space, freedom of self expression, a "let's try attitude". It means a culture with fewer ties and dependencies, with simpler checks and balances, ones that trusts their employees that they are capable to solve problems and learn from mistakes. It means a culture that requires dialogue but abstains itself from being prescriptive. To me, these goals are well met by the airplane model described above.

Initiative is fragile. Push an employee too much to explain himself, to give too much proof that he is right or punish him (even slightly) for small mistakes and the initiative is gone. Next time he will just wait to be told what to do.

I imagine organizations and management lines as meshes of masses connected through springs. The more ties a mass has and the stronger the springs, the lower the movement ability an individual has. And with the lower movement ability, the exponent described above (a function of the ability of the individual to excel in a given context) simply converges to the comfortable average, while the whole system slows down due to lost energy to friction.


Closing thoughts:

The article describes three roughly independent management ideas centered around organization design. There is a red line though. That is systems (of humans) are not to be taken for granted. It is not enough to have the brightest in the room to get results. How you organize them, how you make them communicate (and in humans communication is extremely expensive and tricky due to the low bandwidth of conscientious thought and speak and due to the whole set of unspoken assumptions and emotions involved), how do you setup systems with enough degrees of movement to maintain individual initiative and still have some control over the direction of flow is what gives the higher returns. And all these are layered on a leadership style that must be humble, willing to listen and to point people in the right directions, not for completing tasks, but rather for gathering information and  learning, and then have them take their own decisions with confidence.

Friday, February 17, 2017

On Collaboration - Again

In an organization (or life) there are like three major categories of positioning which you can have towards reaching your objectives. They may be either enforced upon you (by the situation or organization) or they might simply be self enforced by your convictions, confidence or values. These categories broadly are:

1. Asking for permission.
2. Discussing facts, ideas and needs to reach the best possible outcome. Eye to eye.
3. Reaching your goal independently.

For the first option you need to consider three things if the organization enforces it on you. The positive one - what am I missing? Maybe there is a piece of information I don't have access to and asking for permission is a gate aimed at creating an opportunity to learn more (or risk management). This might be the case especially if you are new to the organization or to a project. The other two are not so flattering; it might be a either sign of personal weakness / lack of self confidence (you need permission from a higher moral authority to proceed the way you think it is the best) or a sign of a dysfunctional relationship / organization — political, hierarchical and dependent. Nevertheless, having to ask for permission needs an explanation and an open dialogue. There might be very good reasons behind.

The second option may add significant costs and overhead (think thread contention in software engineering) but opens the doors for better decisions (albeit slower) and, in many cases, a more rewarding job experience and better team bonding. People learn to trust each other in the process of passionate exchange of ideas. They start trusting their ability to think and find rewarding the intense mental process of constructing a better solution together with colleague.

The third is the opposite of the second more or less. Works when the effort to collaborate is higher than the gain obtained from walking the extra mile and when there is very little interdependence. Will encourage silos if practiced too often and will generally result in a degradation of knowledge sharing and communication.

This article is about independent, self confident people, that don’t need to assert their personal influence through the use of their position. This article is about reaching the best possible outcome in the context in which one might not have the full picture. This article encourages people to communicate, share and collaborate. It encourages eye-to-eye dialogue. It opposes dependencies and complexity. It encourages this: https://m.signalvnoise.com/depend-less-on-each-other-507fe0e23e4b#.pg63sff43

Point no 1:

If you don't trust your employees, why hire them in the first place?

The second option is about trust above all. It implicitly asserts: I trust your ability to collaborate and I trust your capacity to think clearly and engage in a constructive conflict of ideas. I trust your capacity to support your view of the world, I trust your capacity to understand the other's view of the world and I trust your capacity to build on several different views. Preferably not a compromise, but a solution that incorporates all needs. A win-win.

Compromise is the lose-lose type; everyone loses, but just a little bit. You get part of your stuff, but pay a price.

Point no 2:

The problem space may be too vast for one person to comprehend. This is why we need teams and collaboration. And collaboration means dialogue, patience, empathy, trust and willingness to take time to reach a win-win arrangement. Again, not a compromise. But again, it is costly, so brace yourself. Sometimes a compromise might just be good enough. But not the first option and, I dare to say, not desirable, as it leaves everyone a bitter taste. Plus it weakens the culture. And I think culture should be first; thus the default solution is always to collaborate.

Point no 3:

Option 1 is not even hierarchy. It is an assertion of discretionary power.

Point no 4:

I understand and respect roles and responsibilities. It is a form of respect towards the people and the organization at large. Their role is to streamline decision making and settle expectations. They create clarity and diverge discussions from trivialities.

I don't understand discretionary power. When I say "please check with me before", I don't mean "ask for permission". I mean, "please let me know because I am interested" or "let's first discuss because I might have some perspective to bring to the solution". This is an invitation to collaboration, it is trust, it is respect. It is "I trust you to have confidence to allow me to intervene and have an eye-to-eye dialogue on this common topic of interest".

Point no 5:

I am not against hierarchy as long as it is based on a common understanding of roles and responsibilities. Sometimes you simply have to say "guys, we're doing it like this" if the discussion reaches a deadlock, the matter is trivial or time is of the essence (it matters more than the quality of the chosen path). Attention: it is addictive, dangerous and will erode culture (ownership, involvement) if not practiced with exquisite care. Better err on the dialogue side.

Point no 6:

Dialogue between people may be very difficult to initiate and will generate resistance and bruises in the beginning. Hard to start, hard for everyone, especially if the culture is not used to it. In a culture of escalations and self-protection, it will lead to a lot of conflicts, nit picking and will probably increase attrition. Treacherous path. But have faith. Time is on your side and your continuous push will change the culture in the end. There will be followers.

Point no 7:

Support the followers of the culture you want to create. Sometimes you might have to be political to be able to move on (and not completely bring the system to a dead halt), but this is time borrowed on the expense of the culture and it will create confusion on where you really stand. There is a price in everything.

Point no 8:

Sometimes collaboration is not about reaching a mutual agreement. You may simply reach out to a peer to brainstorm or seek for a different view of the situation at hand. He may not have a stake in the outcome. Same process applies. Advice is optional to follow, yet asking for advice is the responsible thing to do - the "you don't know you don't know" stuff. In this case win-win is not about winning or losing something. But rather the process of harmonizing two different view points for the sake of reaching deeper insights.