book-notes

Managing Humans: Biting and Humorous Tales of a Software Engineering Manager

by Michael Lopp

The Management Quiver

1. Don’t Be a Prick

A great manager is someone with whom you can make a connection no matter where you sit in the organization chart.
You must see the people who work with you. Constructing an insightful opinion about a person in seconds will make you a phenomenal manager.
It is your full time job to listen to your employees, to mentally document how they are built, and to fulfill their needs.

2. Managers Are Not Evil

The what-do-you-do disconnect between employees and managers is at the heart of why employees don’t trust their managers, or find them to be evil.
Run away from any evil manager, or someone who puts themselves before their team, who lie, and who have no ability to lead.
Because you don’t understand what someone else does at your company, you’re automatically biased against them. And because you understand your job intimately, you believe it’s more important.
Tell your manager what you do and why it matters. If they aren’t an engineer, find a way to speak their language.
A manager’s job is to take the skills that got him promoted, and make them scale. This means building a team to reinforce where he is weak.
Regardless of the relationship with your manager, you’ll speak differently to him than you would to a friend, because he’s part of the organization.
As a manager, you carve out time for regular one-on-ones so that you have a chance to learn.
Delegation is a slippery slope for managers, because pure delegators are slowly becoming irrelevant to their organizations.
People who work for pure delegators don’t rely on them for work because they can’t depend on them for action, which pushes that manager out of the loop.
Politically active managers may be slimy, but they are informed, and they know when change is afoot and what action to take to best represent their organization.
The organization’s view of your manager is its view of you. Judge his clout when interacting with his superiors, or with his cross-functional peers.
Your manager is not a manager until he participated in a layoff, or participates in the constructive deconstruction of an organization.
You want to see who your manager will become because it’s often the first time he sees the organization is bigger than the people.
A successful organization is built of layers of people glued together with managers, who translate between layers in both directions.

3. Stables and Volatiles

The birth of a 1.0 launch initiates the split of a development team into two groups: Stables and Volatiles.
Stables happily work with direction, appreciate plans, calmly assess risk and mitigate failure, and tend to generate process.
Volatiles have issues with authority, seek a thrill in risk, build a lot but nothing stable or beautiful, aren’t reliable, and leave a trail of disruption.
Volatiles turn into Stables by building process and carefully describing how things should be done, because they have the scars and experience.
These new Stables hire people who are familiar, who are predisposed to be Volatiles, which in turn leads to new disruption.
A Stable’s choice of disruption is within the context of the last war, while a second-generation Volatile will remind you “there is no box.”
As a leader, you need to figure out how to let the Volatiles disrupt, while constantly negotiating a temporary peace treaty with the Stables.

4. The Rands Test

A growing group needs to continually invest in new ways to figure out what it is collectively thinking, so everyone knows what’s going on.
When shit hits the fan, don’t cancel your one-on-one with folks who are responsible, or with folks who can prevent future fan hittage.
In a team meeting, kill lies, and identify what is broken and start discussing how to fix it.
The presence of status reports comes down to control, a lack of imagination, and a lack of trust in the organization.
Be comfortable saying “no” to your boss. If you are always on your best behavior and unwilling to speak your mind, then something is wrong.
Independently judge whether your company is growing or dying, and develop a defensible opinion regarding the state of the business.
A regular meeting where everyone can hear the CEO explain his or her vision of the company, and that allows anyone to ask a question, is vital.
Part of a healthy organization isn’t just that information is freely flowing around; it’s that people are leveraging it or acting on it.
Busy feels good, but it is usually tactical and not strategic. Find time in which you’re investing in yourself at work.
In the absence of information, people make shit up, and if they feel threatened, what they make up amplifies their fears. So kill the gossip.

5. How to Run a Meeting

When two people talk, it’s a conversation. With three or more, it’s a meeting, which needs rules so people know when to talk.
An alignment meeting are regular meetings with tactical communication exchanges. Creative meetings require diving into solving a hard problem.
The agenda answers the question of how everyone can get out of a meeting so that they can actually work.
The referee shapes the meeting to meet the requirements of the agenda and the expectations of the participants.
The referee must scan for attendees who aren’t engaged. If someone is doing anything except listening, then they aren’t listening.
To re-engage someone, ask a question relevant to the current state of the topic, referee silence, or change the scenery.
As a referee, own the meeting. Summoning the dictator to shut someone down is a last resort, because then everyone may shut down.
A good referee will improvise, whether letting someone ramble who’s onto something, or cutting a meeting short because progress is blocked.
Meetings must exist, but meetings cannot be seen as the only solution for making progress.

6. The Twinge

Freshman managers think it’s their job to be responsible for their team’s every thought and action, but you must learn to delegate.
You must understand the art of evaluating a Spartan set of data, extracting the truth, and trusting your “Twinges.”
Management is a total career restart, and so approaches you use for products isn’t going to work for people.
As a manager, your day is full of stories. Always be asking if you believe each story, knowing that it’s incomplete, and that it supports one point of view.
Sniffing around pisses people off, and may be interpreted as micromanagement, but it’s drawing on your past experiences to find failures, which cause Twinges.
If you don’t keep a story in check with a Twinge, and that story jumps from one person to the next, then you let a lie propagate.
In summary, listen to stories, map them to your own experience, and ask questions and demand specifics when there is a Twinge.

7. The Update, the Vent, and the Disaster

Your job in a one-on-one is to give the smallest voice a chance to be heard.
Always hold a one-on-one at the same time each week; this sends a weekly reminder that you are there for them, no matter how busy.
Give a one-on-one 30 minutes at least; don’t reduce the time because you have so many people working for you, because you work for them.
Start by asking “how are you?” It’s deliberately vague so that the recipient can’t help but put themselves in the answer.
As you listen to a one-on-one, put the person in one of three buckets: the update, the vent, and the disaster.
A successful one-on-one is not a status report, it is an opportunity to learn something new above the daily grind of business.
If the one-on-one is a status report, listen twice as hard for something you can discuss, investigate, and explore.
If that fails, then either come with prepared points, do a mini-performance review, or discuss your own professional disaster.
Don’t confuse a vent for a conversation. They don’t want a solution, they want to be heard, and you must listen as long as it takes.
A vent concludes and you can jump in when the other person loses steam, or it devolves into a rant.
With a disaster, the worst response is any semblance of emotion. Don’t ask questions, just be quiet and let the emotion pass.
With a disaster, you’re not experiencing the problem anymore, but the employee’s emotional baggage regarding the problem.
A disaster is the result of poor management, when someone believes that the only option for catalyzing change is to totally lose their shit.
A one-on-one is your chance to perform weekly preventative maintenance while also understanding the health of your team.

8. The Monday Freakout

If someone is going to freak out, it’s going to be on a Monday, after fretting over the weekend about the work that awaits.
When the freakout happens, don’t participate. Instead listen and maintain eye contact, and repeatedly nod.
There is likely a very real issue underlying the freakout, after the noisy preamble designed to get your attention.
When you take the reins, ask questions; this moves the person from the emotional state to the rational one.
One pleasant side-effect of attacking freakouts with questions is that the person is already close to a solution, so dig for it.
The fact that a person is screaming at you is a good sign that he clearly, loudly cares. But you still screwed up.

9. Lost in Translation

Beginners are not burdened with the complexity and depth of understanding; they shine brightly with enthusiasm until The Fall.
When getting to know an employee, the first question you want to be able to answer is, “What does this person want?”
When you know where someone wants to be, only then can you start to figure out how to get them there.
When communication is suspect, rely heavily on clarification. When you say something that might be ambiguous, ask “What did you hear?”
In return, when you listen, and the topic or intent isn’t abundantly clear, restate “Okay, what I heard was…”
Sometimes you must verbally go back and forth until a work commitment is stated, because it is never implied.

10. Agenda Detection

An informational meeting simply has talkers and listeners; there is no problem to be solved other than the transmission of information.
At a conflict resolution meeting, some problem needs to be solved, and agenda detection is more complex.
Participants are players, who participate and want something out of the meeting, and pawns, who are silent or instruments of running the meeting.
If you’re sitting in a meeting where you’re unable to identify any players, get the hell out.
The pros are the players who are on the winning side of the issue, while the cons are getting screwed and look pissed off.
A good pro doesn’t acknowledge that they’re the pro, and so they don’t have to take the heat for whatever the conflict is.
To get out, figure out what the cons want, then synthesize everything into constructive next steps and communicate that to the cons.
If you’re 30 minutes into a meeting and can’t figure out what the issue is, then there are too many issues, and it’s time to go.

11. Dissecting the Mandate

There are three distinct phases to the mandate, which is decide, deliver, and deliver (again).
When the debate is no longer productive, and people start confusing the emotion with the decision, then it’s time to make a decision.
For every person who has a strong opinion, there are probably four others who just want someone to make a decision so they can get back to work.
A mandate may annoy the concerned parties, but the silent majority will appreciate the peace and quiet after your verdict.
A good mandate takes moxie. There team has to leave the room knowing a decision has been made, and there is no wiggle room.
If your team has argued for awhile, a mandate feels less like laying down the law, and more like relaying the results of an investigation.
Deliver (again) is individually taking the time to express your reasoning to concerned parties, both winners and losers.
Expect venting from the losers. If instead they’re nodding their head, they don’t believe the battle is over, so you must have them open up.
If you have to relay some higher authority’s mandate, you must figure out its justification to satisfy the rest of your team.

12. Information Starvation

For each piece of information you see, you must correctly determine who on your team needs that information to do their job.
The creation of information is the act of creating context and foundation when there is none.
When you hear gossip, listen not only for what is actually being said, but for what informational gap in knowledge is being filled by this gossip.
Perhaps the biggest loss of essential information is when managers rely on their brains as to-do lists.
Maintain a consistent flow of information. Even if it’s useless to you, you never know who on your team may care about it.
Taking the time to give each piece of information that you’re passing on a bit of your personal context never hurts.
Your team is always going to tell you what they need to know. Employ some aggressive silence to bring it out of them.

13. Subtlety, Subterfuge, and Silence

Management is chess; when presented with a problem, look at the board, figure out the consequences of each move, and then pick one.
Subtlety starts with humility; sometimes your approach needs to start with admitting that you don’t have all the answers.
Subtlety finishes with elegance; you solve the problem in an ingenious, novel way that builds and refines your management aptitude.
Subterfuge is a risk. Using it for good means keeping the intent honest, but it doesn’t mean that someone isn’t going to be pissed.
To talk about something relevant, you’ve got to gather and to process data. In silence, you can assess.
Everyone’s basic agenda is visible after talking to them for 30 seconds. Everyone says what they have and what they need.
Use silence to learn about your coworkers, and to construct a better picture about how to interact with them.

14. Managementese

Managers are hubs of communication. The better they communicate across boundaries, the more data they have, and the better their decisions are.
When you use managementese with an employee, they usually know what you’re talking about, but you’ve self-identified as manager.
Managers in a hurry needs to remember that managementese puts you a few key metaphors from sounding like a used-car salesman.
When you’re talking to individuals, ditch the managementese, and talk to them using the familiar language of a friend.
Your goal is to have a conversation, and so both people sitting at the table need to trust and understand what is being said.

15. You’re Not Listening

The most basic rule of listening is: If they don’t trust you, then they aren’t going to say shit.
A good conversation begins with a bunch of words elegantly connected with listening; it all starts with the ability to listen.
Eye contact is the easiest way to demonstrate your full attention, and it’s also the easiest way to destroy it.
Keep asking stupid questions based on whatever topics until you find an answer where the other person lights up.
Being a curious fool builds connective tissue, allowing you to develop a mental profile of someone, and setting you up for bigger conversations.
To stop on a point, repeat their last sentence by saying “What I hear you saying is…” and repeat your version of their thought.
This communicates that you are directing your full attention to understand what the person said and what it means.
When you can’t find the question, segue, or words to bring out what the other person wants to say, disrupt the conversation with silence.

16. Fred Hates the Off-Site

At some organizational scale, natural cross-pollination and communication activities that used to happen organically can no longer occur.
Off-sites create a space and place where a team can bond, a strategy can be devised, or you can begin an epic journey.
You can’t invite everyone; you must select a group of folks who are going to best represent the company on whatever huge problem you’re solving.
Everyone should present at the offsite. To reduce attendees, cut people who can’t present anything meaningful, or people presenting the same thing.
If you invite someone who is not presenting, then they should speak up randomly and brilliantly.
The off-site is to create grounds to speak heresy, and that’s easier when you aren’t surrounded by visual reminders of obvious constraints.
The Master of Ceremonies is the person responsible for not just moving the day along, but also knowing when to stop and pivot.
The Taker of Notes is tasked with not only capturing the bright ideas, but the right ideas.
Avoid personality tests. They apply clever labels to people, but to really understand one, solve a hard problem together.
Don’t invite external facilitators; they don’t know the culture, the problem at hand, the politics, or the personalities.
An off-site must be at least two days long, letting people soak in a problem overnight, and then attack it the next day.
Unless the energy of an off-site is channeled back into the workspace and immediately acted upon, then an off-site is a frustrating opportunity to dream, but not to act.

17. A Different Kind of DNA

Everyone wants to grow, but in many companies the only perceived growth path is via management.
Job grades are a distraction packaged as a solution to the fact that we don’t have a good idea how to grow engineers outside the management hierarchy.
We need managers to scale responsibility and communication, but we need to dispel the idea that they are the exclusive decision-makers.
DNA, or a design n’ architecture, is a formal meeting with bright engineers from across the team or company tasked with a specific purpose.
It has the best candidates to vet the idea, to talk about how to make it better, to constructively criticize, while being drama- and politics-free.
If you don’t contribute to the DNA meeting, you won’t be invited back. And if you don’t bring your A game, you’ll get mentally trampled.
A DNA meeting is a staff meeting of the influential engineers who don’t want direct reports, but want to lead.
DNA exists as an acknowledgment that a team is led not just by the folks who build the people, but also by the people who build the product.

18. An Engineering Mindset

The first rule of management is to stay flexible.
If you remove yourself from the code, then you remove yourself from the act of creation.
Use the development environment to build the product, so that you understand the language your team uses when talking about getting stuff done.
Be able to draw a detailed architectural diagram of the product, to demonstrate that you understand everything about it.
Write unit tests, fix some bugs, or even own a feature.
By building the product you’re closer to your team, and you’re closer to how software development is constantly changing in your organization.

19. Tear It Down

People who want flat organizations don’t understand how groups of people organize, and haven’t built anything with more than a few individuals.
There are three leaders: The Lead, The Lead of Leads, and The Director.
The Lead is at the beginning of leading the work, not doing it. They are tactical. Their focus is the team.
The Lead of Leads no longer has any hands-on responsibility. They are equal parts tactical and strategic. Their focus is across the company.
The Leads of Leads run the company, because they are the ones who are ensuring that the work actually gets done.
A bad Lead of Lead is fatal; they lose touch with Directors and lack strategic data, or lose touch with Leads and lack tactical data.
The Director curates the vision. They are ideally completely strategic. The Lead of Leads must translate this vision into action.
There is no hierarchy in the roles described above, because leaderships comes from everywhere.

20. Titles Are Toxic

Titles were created as a way to give folks a path toward growth, not for judging someone else’s importance from a business card alone.
A job is a well-defined thing that has a clear and easy-to-understand set of responsibilities, while a title often has neither.
The first growth path is the lead or management track, which is there so that communications and decisions can be sensibly organized.
Your titles might be toxic if they don’t reflect a job that you would consider to be of real and obvious value.
The second growth path, assigning titles to engineers, cannot capture the seemingly infinite ways in which people evolve.
Titles place an absolute professional value on individuals, while the reality is that you are a collection of skills of varying ability.
It is a tall order for a title to capture expected ability, to measure seniority, and to serve as a measure of compensation.

21. Saying No

When the team no longer questions the decisions of a manager, that manager feels like his decisions are always correct, which is statistically impossible.
The good managers are those who have learned how to recover from decisions with dignity, and with help from the team.
Saying no forces an idea to defend itself with facts, and for your manager to stop and think.
Saying no is saying “stop,” and when everyone thrives on movement, the ability to strategically choose when to stop is a sign of a manager willing to defy convention.
Don’t be paralyzed by the fact that you’re one big, bad decision from being out of a job. Embrace the confidence of being “the boss.”
You are responsible for making great decisions, and the best way to do that is to involve as much of your team as possible in those decisions.
By including your team in the decision process and creating an environment where they can say no, you’re creating trust.

Part II: The Process is the Product

22. 1.0

You’re going to be screwed at some point. Keep thinking, don’t yell, treat those you work with decently, and you’ll be fine.
Maslow’s hierarchy of needs is helpful with folks on the edge, and can provide insight when someone is stressing out.
Don’t trust charts and graphs, because they paint the world in a clean and linear fashion that only supports the message of the author.
In the Rands 1.0 hierarchy, the top is the Pitch, or the idea. You’re in a hurry, so don’t forget it.
The People are next, and no one is indispensable. You must let go of people who aren’t working on the 1.0 product with a sense of urgency.
Then Process, which defines communication. The process doesn’t have to be good, but it has to be stuck in a place where everyone can see it.
A great stagnation warning sign during 1.0 is when someone decides to create an organization chart saying “This is who does what.”
Finally, you have Product. You don’t have a company until you have a product, which a neutral party must validate because you’ve gone nuts.
The lower the failure in the pyramid, the higher the cost. A failure of pitch is a structural one that affects your entire company.
What you’re really building in a 1.0 is a lasting, interesting culture that continues to build great products if you’re lucky.

23. The Process Myth

Engineers don’t hate process. They hate process that can’t defend itself.
If you dig into process, you’ll find the circumstances that led to its necessity, how it could be awesome, and your role in maintaining that awesome.
Process is created not as a means of control; it’s being built as documentation of culture and values for The New Guard.
HR is good at defining process, but bad at explaining the culture. Process should be written by those who are also experts in the culture.
When cultural bellwethers leave, so does their cultural context and understanding of the root pain that defined these bulleted lists.
A healthy process is required to stand up to scrutiny, and when a process fails to do so, it must change.

24. How to Start

Beginning has three phases: You’re either fretting about starting, you’re preparing to begin, or you’ve begun.
Stress is a creativity buzz kill. When you’re stressed, you’re in survival mode, but elegant solutions require offense.
Mornings have the gift of optimism because nothing has screwed up your day yet.
Evenings are dark, repetitive reminders that no matter what you do, time is going to pass and you’ve likely wasted some of it.
Mornings allow you to flex the creative side of your brain; evenings, when you’re tired, allow you to flex the logical side.
A hard thing is never done by reading a book or an article about doing it; a hard thing is done by figuring out how to start.

25. Taking Time to Think

The time to kick off deep thinking is right after your last major release, when every lesson of the prior release is in the forefront of the team’s mind.
Schedule a brainstorm meeting and a prototype meeting in the same week, so that no one forgets everything over the weekend.
Assuming you have an idea of what to talk about, invite those with an educated opinion; otherwise invite people chosen at random.
Avoid inviting “obstructionists” who map every new idea against previous experience and then declare the idea “unoriginal.”
Leave the first meeting with five hot topics that people want to address. Create prototypes, wireframes, etc. in the second meeting.
Red flags as weeks pass are constantly revisiting decisions, the same list of attendees, people venting for too long, and the to-do list always growing.
These meetings will slowly die off as you move from design to development. In general, questions should be getting answered, not created.

26. The Value of the Soak

You can spend a lot of energy deciding on what the big decisions might be, but that’s much less important than making the decision.
Active soaks are activities that you can direct and usually require gathering content.
Passive soaks are activities where you point your brain in a random direction and pray.
To do an active soak, ask dumb questions to form a picture, pitch a stranger your mental picture, and then iterate.
To do a passive soak, sleep on it, because your subconscious can construct elegant solutions when you least expect it.

27. Capturing Context

Value is created when people choose to capture context and share the context of their content.
We need our tools to allow us to capture context at the moment when we’re being bright, not Friday at 4pm when we’re escaping work.

28. Trickle Theory

The only source of measurable truth regarding the product is the bug database.
The Critic is the internal voice who does careful and critical analysis of your life.
When you put an item on a to-do list, you avoid conflict with The Critic, who will argue with you about starting an impossible task.
Tasks are on a spectrum of impossibly dull (requiring no mental effort but vast in size) and impossibly hard.
In either case, the first move you must make is starting the task. Progress and momentum yields confidence, silencing The Critic.
After you start working, iterate. Fine-tune your process to eliminate inefficiencies.
If you’re working on an impossibly hard or impossibly dull task and you’re blocked by boredom or confusion, mix it up.
Mixing it up silences The Critic, stimulates your brain with new material, and still allows you to continue processing the task in the background.

29. When the Sky Falls

Given a disaster, your first goal is to understand absolutely everything you need to know about its current state.
Get a War Room to break everyone from their current flow. Focus on breadth of information acquisition until you have a glimpse of a theory.
Vet your theory with at least three qualified others who are not directly involved in the preceding step.
Once your theory is vetted, put it on the whiteboard and assign owners to each and every task. Don’t let any name be yours.
While others work, your job is internal public relations. Communicating the status is also a way of vetting the plan and the progress.
When the sky is falling, fixing the situation is a bandage, but understanding what you’re truly trying to fix is the cure.

30. Hacking is Important

Hacking creates new things which is a disruptive act, which scares the reasonable people who represent the majority.
Those who are responsible for maintaining and building on success will not understand why hacking is important.
A healthy product company is at odds with itself, as it must normalize to create predictability, but also build something new to disrupt that normality.
Failure to create predictability results in chaos; failure to hack and create chaos results in losing to more agile competitors.

31. Entropy Crushers

A project manager is responsible for shipping a product, whereas a product manager is responsible for making sure the right product is shipped.
A program manager is an uber-mutated combination of both that handles multiple interrelated projects like, say, an operating system.
Each new person on your team increases the cost of communication of ideas, making decisions, and detecting and fixing errors.
If you’re a full-time engineering manager of growing team, but you’re also serving as a project manager, then you’re half-assing one of those jobs.
A good project manager will measure, control, and crush entropy. They own execution of the machine ensuring that everything gets done.
A project manager will ask upon arriving on the scene “What the fuck is going on?” and then create artifacts of insight.
What your team, and your culture, needs out of a project manager depends on the people, the team, the culture, the projects, and this moment in time.
Fire any project manager who goes crazy with power and becomes political, using information to control rather than illuminate.
The arrival of crap project managers is that you’re punishing inefficiency with useless bureaucracy, which creates more inefficiency.

Part III: Versions of You

32. Bored People Quit

Boredom is not initially catastrophic. It shows up quietly and appears to pose no immediate threat, making it both easy to address and easy to ignore.
To detect boredom, look for any change in their daily routine, ask if your employee is bored, or they’ll tell you and you listen.
Every second someone is bored, a second passes on an internal clock, and after some amount of time he or she gives up and quits.
For each person on your team you must be able to answer: Where are they going? And what are you currently doing to get them there?
If someone doesn’t have a project that makes them light up, let them experiment. Your job isn’t just building product, but building people.
Dole out shit work fairly. Be aware of who’s doing it, communicate that you’re aware, and tell them when they’re going to be done.
Promising productive and creative time that is only taken away by urgent tasks only accelerates the boredom clock.
Remove daily distractions that pull people away from their work. These are more costly than they appear from the overhead of context switching.
As a manager, don’t stop coding, otherwise you’ll have a harder time talking to engineers because you’ll forget how they think and how they become bored.