Hacker News Reader

Comment by ballenf | Hacker News Reader

Making a big mistake as a manager can cost the company a lot more than a big mistake as an individual Engineer, Architect or Developer, especially if you consider that the latter roles will catch many of each others' mistakes.

The manager level can result in entire teams quitting, failing to support each other or an atmosphere where catching each others' mistakes is discouraged.

Some jobs are paid based (roughly) on the value they bring and others are paid based on the threat/cost of large mistakes.

View on HN

This is factually not true. Developers can make big mistakes to bring entire companies down with them. Of course it depends on the definition of "mistake". However I can claim that a single defect caused by a developer can cost the company money and downtime, if nothing else a bad user experience.

Managers don't generate defects. Developers do. Not to mention that a lot of the times it's much harder to reverse an engineering decision vs a managerial one.

Most outages are caused by developers and IT, not managers. This perception that there is higher risk as a manager is false. As a developer I make multiple decisions daily, interpret vague requirements, make judgement calls, request new technologies etc. My manager understands about 20% of the decisions I make. Here is the kicker since I work in DevOps, with a super heavy emphasis on the Dev side. 70% of my decisions have wide ranging effect, much wider than my manager's eight person team.

Finally note that no one really talks about 10x Managers, but people talk about 10x Developers. Why is that? Managers are glorified cheerleaders. Most managers are average, their value is vague but they are controlled by managers so its an endless loop till you hit the worker layer that get things done.

You sound like someone who has never been a manager or has no idea what a manager does.

Yes, you're right that defects are made by developers. But you'd be hard pressed to find an example of a software defect that destroyed a company. Even if you as a DevOps engineer cause your company to be down for a day, it probably would have little effect on your company's bottom line.

I was in charge of Reliability for Netflix. Our whole site was down for Christmas Day one year. No one was fired, and it barely affected anything. That was a pretty major defect.

However, the way that management handled that outage had a huge effect on the company surviving that incident. Making sure we focused on the right problems to solve to prevent it from happening again. Creating entire new groups of engineers and then hiring them to solve specific problems of moving traffic between regions.

So yeah, management is really important, and a good manager is definitely more effective than a single engineer. And management decisions can destroy a company, but I doubt you can find even one example of a single engineer's decision that took down a company.

I was a Sr Manager for about 4 years, two different "teams", one IT based, one QA based. Regardless, perhaps we need to define the word "Manager", mainly because your example said:

> Creating entire new groups of engineers and then hiring them to solve specific problems of moving traffic between regions.

Albeit admirable, like heck yeah that was a great outcome, it's not a manager decision. Most likely it was a VP or Sr Director decision. Perhaps a manager brought it up as an idea. So if we are talking about Sr Director's+ then yes, their decisions can absolutely bring down a business. Granted I am not familiar with Netflix's org chart so perhaps it works differently, from some basic reading it looks like Netflix follows a decentralized command of sorts.

My beef with the original poster was the implication of risk. Both managers and devs have risk with their decisions. It's not just a manager thing. In the end, when Netflix went down where did the mistake come from? A manager or an architecture decision? Probably it was both.

> In the end, when Netflix went down where did the mistake come from? A manager or an architecture decision? Probably it was both.

In that particular case it was an engineer making a mistake, but my point was that it wasn't anywhere close to a fatal mistake.

But more importantly, when a team had repeated technical failures, it was usually the manager who suffered the consequences. Especially if their failures were causing problems with other teams.

The manager had the higher risk job because they were blamed for their team's failures. Therefore they had to be compensated to account for that risk to their job.

It's the same reason CEOs get huge golden parachutes. Because their job is constantly at risk for not only their decisions but the decisions of everyone below them. So they get compensated for that risk.

Why does potential for costly mistakes result in greater perceived value and higher pay?

Higher risk begets higher premiums.

Why?

The relationship between contribution and compensation is obvious. If you bring more value to the company, the company compensates you directly with a portion of that value.

But why does a position with greater risk necessitate greater compensation?

EDIT: Everything below here was in the original post. However, I re-read it and realized it had little to do with the relationship between risk and compensation.

I don't buy the idea that managers are liable for the mistakes of their reports. That would cause a snowball effect in large organizations, resulting in extremely high turnover for leadership positions.

Managers share some liability for their reports' mistakes, but they also share some reward for their reports' successes. Those liabilities and rewards are split across all of their reports, resulting in what are usually relatively stable positions within a company. A lazy manager can get along by just regularly pruning reports who produce more liability than reward.

> But why does a position with greater risk necessitate greater compensation?

There are various methods to address risk:

1. Avoid the risk

2. Reduce the risk

3. Transfer the risk

4. Accept the risk

When you buy insurance, you're doing #3. If you self-insure, you're doing #4. Either strategy requires putting some money towards premiums, the expected possible cost of the risk.

It's much cheaper to pursue #1 or #2. You're then saving an amount exactly equal to the premiums don't have to pay under #3 or #4.

Thus, you can pay a person to do #1/#2 a fraction of the equivalent cost of #3/#4 and still come out ahead.

> A lazy manager can get along by just regularly pruning reports who produce more liability than reward.

Yup, and the books will balance.

It’s a generally accepted principle in economics/finance that the more risk you take on, the more up-front compensation you can expect. This is less about the risk the reports pose to the manager and more about the risk of the manager position itself and the high opportunity for spectacular failure.

> Making a big mistake as a manager can cost the company a lot more than a big mistake as an individual Engineer, Architect or Developer, especially if you consider that the latter roles will catch many of each others' mistakes.

Of course anything is possible, but there is often a chain of managerial command (managers of managers) to catch mistakes. It's pretty hard for a single, middling manager to inflict costly mistakes anymore than an individual engineer. Also, I've seen management levels that are 0 - 1 subordinates deep (managers of none).

Reply on news.ycombinator.com