4 ms·

I'm a developer. I dare you to provide a metric that can measure my outcome. There is no metric that can measure my outcome.

wins32767 8y ago
Let me ask you this, can you tell which of your coworkers are good developers or a bad ones?
- shikoba 8y ago
  So your metric is just an opinion?
- skybrian 8y ago
  Not unless it's an extreme case. I found writing performance reviews excruciating because I don't pay that much attention to what other people do and hate to judge people.
b_tterc_p 8y ago
A combination of self assessed competence and average actual performance vs expectations in terms of oversight required, time to complete required, lead time on flagging delays, effectiveness on coaching junior members, effectiveness of upwards management, pleasantness to work with
- chopin 8y ago
  All this can be expressed in cold, hard, and comparable numbers? I very much doubt that. Any system which wants to measure these traits is doomed to fail.
  - b_tterc_p 8y ago
    Some of them can. Some of them can be measured in soft feedback ratings on simple scales.
    - chopin 8y ago
      That's why I said comparable. Soft feedback is not comparable.
  - dr_dshiv 8y ago
    Rules of measurement: 1. Anything can be measured 2. All measurements have error
hopler 8y ago
Why do you do the work they pay you for, if no one can measure it? Seems you've got a sweet deal, get paid to do whatever you want to want regardless of result.
- shikoba 8y ago
  That's how it works with every developers. They could make an opinion about my work, but that's all.
bluGill 8y ago
$: I can measure how many dollars in sales I get after releasing your work. Of course you probably work on a team so I can't say how much of those dollars come from you...
yen223 8y ago
The only way your outcome is unmeasurable is if your outcome had no observable impact. If your presence had any impact, then it can be observed in some quantifiable way. Perhaps it's an increase in customer satisfaction, or a bump in revenue, or an increase in team velocity. All of this can be quantified in some form or another.
- rifung 8y ago
  > The only way your outcome is unmeasurable is if your outcome had no observable impact. I don't think this is correct? Your work having observable impact and being able to show exactly what observable impact your work had in a quantifiable manner are two very different things. I feel this is a huge problem at Google, where performance reviews/promotions are supposed to be based on objective measures. If you fix a lot of bugs, the end user might be much happier. However, to quantify exactly happy you need to have some existing infrastructure, and even if you do have that infrastructure, how do you know whether they are happy because of the bugs you fixed or the bugs someone else fixed? It may be possible but is not really worth the effort for a company to invest in. So, what ends up happening is people (who care about perf/promotions) fight for the projects which will give them the most easily measurable impact, and ignore other things which may well be very important to a project or company but are difficult to measure. Ideally, your performance should only be rated on things that were actually in your control. Admittedly, I don't know how one would do this though.

wins32767 8y ago

Let me ask you this, can you tell which of your coworkers are good developers or a bad ones?

shikoba 8y ago
So your metric is just an opinion?
skybrian 8y ago
Not unless it's an extreme case. I found writing performance reviews excruciating because I don't pay that much attention to what other people do and hate to judge people.

b_tterc_p 8y ago

A combination of self assessed competence and average actual performance vs expectations in terms of oversight required, time to complete required, lead time on flagging delays, effectiveness on coaching junior members, effectiveness of upwards management, pleasantness to work with

chopin 8y ago
All this can be expressed in cold, hard, and comparable numbers? I very much doubt that. Any system which wants to measure these traits is doomed to fail.
- b_tterc_p 8y ago
  Some of them can. Some of them can be measured in soft feedback ratings on simple scales.
  - chopin 8y ago
    That's why I said comparable. Soft feedback is not comparable.
- dr_dshiv 8y ago
  Rules of measurement: 1. Anything can be measured 2. All measurements have error

hopler 8y ago

Why do you do the work they pay you for, if no one can measure it? Seems you've got a sweet deal, get paid to do whatever you want to want regardless of result.

shikoba 8y ago
That's how it works with every developers. They could make an opinion about my work, but that's all.

bluGill 8y ago

$: I can measure how many dollars in sales I get after releasing your work. Of course you probably work on a team so I can't say how much of those dollars come from you...

yen223 8y ago

The only way your outcome is unmeasurable is if your outcome had no observable impact. If your presence had any impact, then it can be observed in some quantifiable way. Perhaps it's an increase in customer satisfaction, or a bump in revenue, or an increase in team velocity. All of this can be quantified in some form or another.

rifung 8y ago
> The only way your outcome is unmeasurable is if your outcome had no observable impact. I don't think this is correct? Your work having observable impact and being able to show exactly what observable impact your work had in a quantifiable manner are two very different things. I feel this is a huge problem at Google, where performance reviews/promotions are supposed to be based on objective measures. If you fix a lot of bugs, the end user might be much happier. However, to quantify exactly happy you need to have some existing infrastructure, and even if you do have that infrastructure, how do you know whether they are happy because of the bugs you fixed or the bugs someone else fixed? It may be possible but is not really worth the effort for a company to invest in. So, what ends up happening is people (who care about perf/promotions) fight for the projects which will give them the most easily measurable impact, and ignore other things which may well be very important to a project or company but are difficult to measure. Ideally, your performance should only be rated on things that were actually in your control. Admittedly, I don't know how one would do this though.