Leadership: Random Notes
Thoughts, questions, ponderings about leadership. Some may be useful. Some may be plain wrong. Anyhow, I need to write these somewhere. Here it is:
- The common pain felt by new leaders is the inability to delegate.
- Alice, the boss, tells Bob, a rising star, to x10 himself by giving him subordinates. Naturally, delegation is his self-multiplication function. Bob is finishing 1 story-points per hour, so Alice expects to have Bob and his three subordinates finish 4 story-points per hour.
- Unfortunately, Bob can't delegate. Bob doesn't trust people to do his work, and not for a no good reason -- Bob's code quality bar is high!
- The inability to delegate often comes from a lack of trust.
- Trust, how do you establish it in tech?
- In the tech industry, the more complex the problem, the more akin it is to art, as in, a huge percent of the work is subjective. There is just too much disagreement to turn to conventions. One can only have 24 hours a day.
- Subjectivity includes, but is not limited to, 1.) how elegant a code can be, 2.) what should the order of priority between correctness, performance, security, and user experience be, 3.) OO-style vs. F-style, and etcetera and etcetera
- (Seriously, modern-day Object Oriented Programming is not even what the visionary Alan Kay originally termed).
- Again, how do you expect to establish trust in an industry that is like art? Can it be simply done by getting to know your teammates? What if, after spending much time getting to know your teammates, it turns out you don't fit each other?
- Taking a step back! How much of leadership is delegation anyway?
- How much of leadership is about management? How much of it is about influencing people to work toward your cause?
- When compared, management feels more tedious than influencing people. Talking about the required amount of work to achieve, managing subordinates feels
O(n)-ish where n is the number of subordinates. Meanwhile, influencing people feels about
ais directly proportional to the number of subordinates and
bis directly proportional to the leader's output, or something along that line. Did we just talk about the two opposite sides of a coin?
- Up until this point, we were only talking about managing downwards, managing subordinates, thinking about them... thinking for them. But there is this page I sometimes frequents for wisdom, saying "A good manager is one who manages in strict priority order; self, superiors, peers and only then subordinates."
- Let's try to do some calculations. Let's have a mock organization. This organization is in the shape of a tree, with 3 layers. A node represents a person. Each node except the bottom has three children. Each node, to be fulfilled, must receive one unit of management effort. About what does a unit of management effort represent, we will talk about it later.
In only self and downward management schema, a person receives units of management effort from their superior and themselves. Let's say the ratio between effort-from-self and effort-from-superior is 1:1 for simplicity's sake. The result is, the minimum number of units exerted by a node is 0.5 while the maximum is 2.5.
In only self and upward management schema, a person receives units of management effort from their subordinates and themselves. Let's say the ratio is the same, 1:1, it turns out that while the minimum, exerted by a node is the same, is 0.5, the maximum is only 1.17, lower than the half of the maximum of only self and downward schema.
- By balancing the exerted effort distribution between layers, we can adjust how many units stream upward and how many downward. A balanced distribution will almost entirely consist of upward streams, with a little portion of downward streams.
- The number is saying something, but these arbitrarily defined values start to sound like pseudo-science (and the dreaded but satisfying graph theory class at university). So let us define what is a unit of management effort?
- If I define a unit of management effort as the amount of service done, it will make the top-level-leader in the upward schema an authoritarian spoiled ruler, and in the downward schema, a spineless ruler. So I'll go with a smoother definition, units of management is the amount of thought given regarding the receiver's problem.
- This definition is good in dominant upward streams. If the leader's concern includes their subordinates' wellness, it's a "help me help you" type of delegation. Either way, what you need to do in this schema is to take a portion of what your superior needs to think for. You work on that while working on making yourself better. Tell your superior what you need, and more importantly, how can your superior provide for you, so they don't have to think! Also, establish trust in your subordinates. Trust, but verify!
- Math checks out (kind of). But do you think this is a good, balanced flow of "thinking about self, superior, and subordinates"? Is the model accurate enough, or does it miss the mark by a lot? Do you have a better way to model it? Or would you approach it differently? It is a very simple model. There are a lot of things in real life that is not represented here, such as managing peers, the need for a different amount of effort from a different person, a maximum amount of attention a person can exert for others (one can only have 24 hours a day after all), etcetera.
- Also, random: upward schema looks like Blitzkrieg.
- Next topic! "I need to take care of myself first!" is a good principle. "I need to take care of only myself!" is a bad principle for an organization. How does one protect an organization from members in chaos and selfishness?
- First option, reward. Reward is easy, either make an agreement to pay someone for their service! Reward can also be a surprise. By giving someone an unexpected reward, the receiver will most likely give back in the form of appreciation and gratitude.
- One thing to note, as an employer, you are not free enough with rewards, especially monetary ones. After office hour, you come over to your teammate that just committed a clean elegant code, then give him money, PLEASE DON'T DO THAT! It is awkward.
- Second option, in which you don't need to give a reward. Give a speech on noble and common goals, rally for good cause, gain supporters, gain masses, think like a politician. You're in tech, but that does not mean you should max out your intelligence and ignore your charisma, no, you need to accrue social currency. Influence, you'll need that, for, like, RewriteXInRust™!
- Is an influencer a leader?
- Now, what happens if influencers in an organization direct people for their gain? Resource, resource is finite! If you don't work together and pursue your agenda, you'll dry up and wilt. A leader is an influencer that works toward common goals for common benefits.
- Remember what happens to greedy leaders/rulers who lost the support of the people! Vive la révolution!
- Also, IMPORTANT! The side effect of uncontrolled influence and power play, subversion and power of the mass, is dangerous! Be careful, huge power may be directed to innocent people too, or worse, to you. A reminder to those who have a healthy working environment, keep it healthy at all costs and be grateful. Once it is corrupt, it's hard to turn back. Keep your friends close...!
- Third option, RULE. WITH. FEAR! (evil laugh).
- Seriously, fear of
dyingdeadline is effective in moderation. Timebox!
- Punishment too, but, you know, use it sparingly. It's not good for your conscience.
- Have you figured out how do we trust someone in an art-like industry such as ours?
- (That question is not me read-baiting you, I'm asking sincerely. I really don't know.)
- You have order in your rank. What does a leader do? What else but move the organization for a cause!
- As a leader, I think it is an absolute must that you know where you are going, have at least a set of principles, have a cause, a destination. You need to have vision! That is only what I thought though, what do you think about leaders having a vision? What if your job is only as boring as accepting specifications and implementing them?
- I think vision is crucial to guide yourself and others. Moving is crucial to keep our bodies healthy. Constant improvement! It keeps morale high.
- So, other than being a manager, an influencer, you also need to be a warlord, a commander. Vision keeps morale high. You need vision in the future to anchor too. You'll need those at least to survive the next biweekly sprint!
- Be confident with your cause and your vision! If you don't believe them, people don't believe them.
- What is a team without morale? What do bored and unmotivated people do? Desertion? Mutiny?
- A neat trick if you and your team have a boring job. Declare war to something. Make enemies! Bad code? Enemy! Regression? Enemy! Bad management? Enemy! Inefficient space/time complexity? Enemy!
- It is important to keep the goal not just technical but emotional. It must be felt, not just dictated.
- Related: I'm not a citizen of USA, but every election, the intensity is way off the chart. It is felt around the world (or at least on Twitter). My guess -- It is because they only have two parties. People concentrate to attack the opposing force at full might because they see only one enemy. Compare it to an election my country used to have -- 40+ parties! Not as fun.
- Rousing a bunch of people requires a bit of talent. I had this colleague who can make war speech enough to make people excited to work. This friend is very bright, technically capable, and has a great taste in design and architecture, but those cannot even compare to his charm!
- Appreciation boosts morale, but appreciation from a superior is only half the story. A superior must be able to make their subordinates appreciate themselves. People who feel to have made a significant change will keep making a more significant one. Involve them!
- What do you think about a leader having excellent technical skills? How would they keep up with the ever-moving technologies? Are they important? HBR thinks so https://hbr.org/2016/12/if-your-boss-could-do-your-job-youre-more-likely-to-be-happy-at-work.html, they're good for morale.
- As a leader on the top who is not doing technical stuff on a daily basis, it's hard to keep up with technical skills. It is a sad truth.
- Technical stuff is the reality. At the end of the day they are all a bunch of data moving, mashing, matrix multiplying. You might be good at the past, technically. But your knowledge might be a bit rusty. Do not lose your grip to reality.
- Ask your subordinates about what's new, what's exciting, what they think is the next great framework. Ask them to update you and your team, and don't forget to tell about that your superior too. Also, you have more experience, can make judgment whether those are actually a good product or just the community being in the honeymoon phase.
- Pride is a dangerous thing. If you are prideful, keep it in check. No one wants a leader isolated in an ivory tower.
- "I told you so" is countered by "how would you approach this back then?", not by anger and resentment. Ask what they would do! If they back down, it means they are probably a jerk. If they are eager and constructive, they actually care. Do not mistake the perception of aggression from them with your pride and insecurity. Instead, always seek what is right.
- All in all:
- 1_ Try to be the best in technical skills, because it makes you a good leader!
- 2_ Don't be too prideful in your technical skills. Don't assume you are the best. That will make you prideful. That will stop you from learning.
- Speaking of being at the top. Aside from having it hard about technical skills, you also have it hard controlling things. You know it is good to have control, but not too much. It is good to be able to trust people with what you want to achieve, and they figure out how to do that. But losing control because you are high up, is not comfortable.
- How do you keep balanced control over things far away?
- Organization values and principles set them up. Find one that strikes people's hearts. "Give back to the community". "Self-improvement always". "Blameless postmortem". Etcetera, etcetera.
- Create a set of protocol(!). It is more strict than principles.
- Last, from me: espionage. If you have someone flexible that you trust and have good chemistry with you, you can install them to where you don't have sight over. They would be a sleeper agent, doing their own stuff, reporting to you once in a while. You can actually do this as a subordinate, install a friend in another team far away to see what they are doing, steal some technology (as long as it is legal and within the organization's rules), for science!
- Last, I beg you, if you are a leader, realize that you have power. You may not feel it at times but it is there. The higher you're up there, the bigger your power. If you are paying for your subordinates, it's a big multiplier.
- If you are in this position, relearn to follow! Reminisce the time you are following people. You will notice that, what you casually is your subordinates' mandate. Choose the right situation when you need to be assertive, and when you need to be a little subtler. If you don't mean what you say to be an immediate demand, make it clear.
That's it. Tomorrow those ideas might change, but now I stand by these. Debate me, if you're feeling constructive about it. Also, try to guess which point is real and which are jokes?
Inspirations and interesting links: