Hacker News Reader

Comment by syshum | Hacker News Reader

Combined with that, many employee's know how to work the metrics to make themselves look more productive than they really are.

So you end up with work being done to fit the metrics, not necessarily work that is of highest need or more impactful to the organization.

I have seen plenty of bad metrics lead to managers no understanding why their dept's appear to be "hitting all the numbers" but also no performing in the way they desire at the same time

View on HN

After my performance review, I talked to my boss and my boss's boss, and asked them if that's really what they wanted me to do, ie. fulfill these arbitrary metrics or if they wanted me to work on whatever made the team better. They said no, of course not, but a good engineer should be able to fulfill those metrics and make the company better. I explained the situation, that the ops person left and they haven't hired anyone to take over, so I'm left holding the bag and they refused to acknowledge it, so I quit.

As a manager myself, you have terrible managers. They should be able to acknowledge the facts of the situation and not put the blame on you. To the opposite, good work, such as taking over something critical when someone left the team should have been rewarded. Metrics are useful, but to only rely on them is what lazy or incompetent managers do.

What was the point of saying all of the above? In the real world there's many bad managers and new managers are assigned to existing projects all the time. Any tool which provides more misleading metrics is bad news.

Proving that the initial problem can be reduced to a people problem is useless.

my guess is they're trying to say that the person was in the right, their managers in the wrong (and they are a manager themselves) and if they run into the situation again, they know what to look for. Also, in the event this person is beating themselves up over the situation (thinking this is the norm, thinking that maybe they made the wrong decision, etc.) it would appear it is intended to put them at ease that they did the smart thing and it isn't like this on every team.

The point is that some metrics are useful tools, and that some managers are tools.

To be more specific about metrics, something that measures what your users care about, or is beneficial to your company is actually useful. Sometimes the most important metric is a zero or one, which is did you ship? Measuring commit frequency seems intrinsically useless.

Why not just ignore the DevOp work and focus on what counts. If no one is measuring ops then it doesn't matter. Someone will get upset the issue will go above you and they will magically hire someone.

If you are covering up for something (no staff) than you will look bad. Push the problem up. Make the people above you look bad instead of taking every bullet.

> you end up with work being done to fit the metrics, not necessarily work that is of highest need

Obligatory reference to (generalised) Goodhart's Law:

"When a measure becomes a target, it ceases to be a good measure."

https://en.wikipedia.org/wiki/Goodhart's_law

Reply on news.ycombinator.com