Hacker News Reader

Comment by bcherny | Hacker News Reader

This article reads like this person has not worked in a well run engineering organization.

In a good org:

- Managers do not set goals. Teams set goals (EM, TL, PM, DS, etc. need to collaborate on the goal and be bought into it).

- There is a system of checks to make sure goals are good. Teams need to make sure their leadership is bought into their goals, and that the goals ladder up to company goals.

- Managers do not prescribe technical approaches. This is something TLs should do.

- TLs are empowered to propose and work on infra and tech debt that are harder to measure with KPIs. These still need milestones and some measure of success.

View on HN

I might misread you. I would believe the best approach is a top-down / bottom-up approach that combines strategic directives with tactical and local goals. Otherwise, you would essentially say that the owners of a company would not be allowed to shape the direction of their own company, but that they would have to accept to be convinced by whatever the individual teams come up with. There are many valid goals that will never make sense for the individual teams to prioritize on their own but require a mix of top-down / bottom-up goal setting

I agree. From my comment:

> Teams need to make sure their leadership is bought into their goals, and that the goals ladder up to company goals.

> reads like this person has not worked in a well run engineering organization.

Could be, it's not unusual. You might be surprised how common it is. And it is useful to address the real world of that.

For sure. I hope it helps others to understand how an eng org should be run, so that they can help adjust their org’s approach. If incentives and process aren’t set up right, there’s only so much you can do to bandaid over the dysfunction.

Wait, so if in well-run engineering shops, the managers don't do any of the things you listed, then what do the actually do?

Generally, managers work on ensuring the business keeps surviving. That’s their main responsibility. Staffing, retention, alignment of sales/ development, etc.

Most good management does that well in harmony with the product. Bad management focuses on that at the expense of the product.

A manager’s job is to make themselves obsolete. This means hiring, retention, mentorship, training, sharing context, etc. until the team is self-sufficient.

Engineering managements job is to make sure the whole endeavor of building and maintaining the software is successful. In order to do this they must wield some power as necessary, but if everything is running smoothly as described above then it’s demotivating for them to be stepping on the toes of the cross-functional ICs (Eng, PM, DS).

This might make EM seem like a cushy job, but in practice things are always a mess at various levels, and a good EM needs to wade in and sort out all kinds of conflicts that can lead to entire teams outputs being wasted. To do a good job the EM must have solid technical death, good product and business understanding, and excellent listening and communication skills. More specific needs depend on the company, org and situation in question. Being a good engineering manager is incredibly difficult in large orgs, and it’s rare to find them giving the difficulty in assessing them and the resulting market for lemons making it viable career path to fail upwards (or at least sideways).

Plenty of companies have the roles of technical leads and managers overlap which honestly can make a lot of sense depending of how your engineering org is setup.

Reply on news.ycombinator.com