Notes on People Management

Why do you need people management?

Let’s think from first principles. Suppose you have a vision, about a certain shape of a product that is wanted by some people. It involves many functions pushing together to make it into a stable business. The two most fundamental ones are, Product/Engineering that puts in the money to build it, and Sales that gets money from selling it - creating a virtuous loop. Raising money from equity or debt gives the former more lead time, but eventually, the product has to sell. As the business grows, you improve the effectiveness of this loop by adding Customer Success that keeps customers engaged, Marketing that reaches more customers, Legal that protects the business, and Finance that does the accounting, budgeting, and making the business more appealing to investors. There are more support functions like HR, Security, Compliance that may get different emphases in different industries.

Someone needs to be directly responsible for all of the functions. This is the role of the CEO. The CEO and the function heads constitute the executive suite - the first team. All members of the exec suite need to think strategically, and be aligned with maximal information bandwidth. Since they more or less self-organize, my notion of people management isn’t primarily focused on this level.

In this post, I would like to focus more on people management inside of each function, with a particular emphasis on Engineering. Typically, the finer the granularity of responsibilities, the less direct the link between individual performance and overall company success. So the performance evaluation gauges, practices, and culture find other anchors to lean on, e.g. engineering principles, rigor, and teamwork.

Let’s go from an Engineering team of one. This usually is frontend, backend, data loading/migrations, infra, CI/CD together. You start with platforms and tools that automate most of them, and instead focus on the core business logic, the special sauce. As business needs become more complex and bigger scale, and reliability requirements become more demanding, it naturally overwhelms the one-person capacity. It’s time to hire out of the situation.

Assuming each person is of similar intellect, you can aim for a best case of Nx factor to the one-person raw productivity, N being the number of team members. This is linear scaling. Of course, if the team is coordinated well, the business impact can be a lot higher than that, because high execution velocity realizes the vision/strategy. But coordination has communication cost, so in reality, scaling people is usually sub-linear in terms of raw productivity. It’s OK to be sub-linear, because again, businesses demand speed of execution, so as long as more people don’t lead to negative productivity, it is worth it for a high-growth company to keep hiring. This is the business/macro-level consideration.

On the individual/micro-level, a manager needs to optimize towards a higher scaling factor, by focusing on prioritization, unblocking, collaboration, and fostering a culture of learning, so team members can skill up. They also need to identify, convince, and hire more talented/experienced candidates onto their team within the budget, to boost up the per-capita productivity. Together, the team’s output is maximized. The management quality is determined by these two axes, in how strongly they can guarantee their team’s high productivity and success within the budget.

A manager needs to develop an active personal loop with every team member, from when they were still a candidate before hire. The idea is, you need to understand their capabilities and drive, so that you can maintain alignment and let them push themselves the most. You need to mature yourself faster than the team, so you could offer guidance, mentorship, and make the right calls. This creates mutual respect and trust, a critical component in leadership dynamics.

When getting new people onboarded, a manager needs to foster opportunities for them to absorb the instutitional knowledge. Talk to them often, and encourage them to sync up with existing team members. A good start forms the baseline expectations. Let high-bandwidth communication be part of that baseline.

A manager needs to make the productivity pressure gently present as a background hum. This is the glue between the airy “team culture” and formal annual evals. In-office work, open floor plan, etc. have this by default - you see how everyone works. Remote work, on the other hand, loses this baseline, and it takes a lot of brushing up to maintain it without getting tedious. It tends to slip into a “nothing happens every day and it’s ok” kind of complacence. You look away, people stop working. Quarterly meetups, proactively showing work progress every day without awkward standups, keeping the communication channels active and by default open can all help. Remote work is great for discovering great talent, but it is inherently disadvantaged in this regard, therefore best only for highly self-motivated people.

Moreover, all managers in adjacent teams have to apply similar productivity pressure, otherwise it creates unfairness and cynicism. If the infra team is driven hard while other user-facing teams are stalling, it makes everyone doubt company priorities. All managers need to feel the pressure in a stronger way than their team members, and drive them towards progress. Otherwise, the productivity pressure effectively does not exist.

There is a very human aspect in this management process. A manager’s effectiveness fundamentally depends on if they can rebuild the team in their vision. A limited version of this is, if they see a deep problem with someone they manage, they need to be able to find someone else who can do a better and adequate job, so they can uphold the productivity guarantee needed by the business. A strong manager can project this position and move the team forward.

You want to avoid the team from getting into a state, where there are too many things going on, and too few hands to get through them. To draw an analogy from computing, too many threads on too few processors cause thrashing. The context switching overhead becomes too big, and you lose the ability to fairly evaluate your team members’ work. Everyone gets a plausible excuse of “too many things”, and very little ends up getting done, and apparently no one has dropped the ball. Without a reliable signal and feedback, the team becomes effectively unmanaged. Sometimes this is inevitable, and you need to deepen the trust between yourself and team members. However, you need to consider scaling the team when there is a chance, and push this signal upward to your leadership.

You also want to avoid having only one person in charge of an entire field of work, without enough trust-building. Absent a convincing backup option to hold the bar, you are at the whim of that one person, for better or worse. It turns an employment dynamic into a partnership dynamic. But the formal power and employment structure don’t reflect that, so you will get struggles and disappointments. This is a common dilemma, and I’m sure it will resonate with a few readers. There is no easy way to get out of this, and the risk is high. The fundamental solution is developing a deeper understanding of the business, the bar (how things ought to be), the industry consensus, your options in the talent market, and the raw human nature. Being able to make a clear case to your team members, your peers, and your leadership allows you to push forward.

To conclude, people management is ultimately about making the business’s productivity guarantee credible. That means hiring, unblocking, and developing people proactively, and being willing to reshape the team when needed. The hardest situations - thrashing or a single “indispensable” owner - force you to rely on a deeper understanding of everything around you, and human nature. That understanding is what lets you move forward when formal authority alone isn’t enough.