3 ms·

The definition of "results" and the criteria by which they are measured is the critical thing. If you measure everyone in a company by how much money they dire

The definition of "results" and the criteria by which they are measured is the critical thing. If you measure everyone in a company by how much money they directly make, then salespeople are the best, and everyone else is just drifting along on their wake. You could measure by functionality added, defects found, bugs fixed, projects completed, requirements gathered, and in each case, you assign merit to one class and not to others. Even if you have a reliable way to measure merit in one job family, how do you align the merit metrics across them?

AstralStorm 8y ago
Trying to turn running a software project or a company into an algorithm is a definition of bureaucracy. You do not look at metrics but judge performance. The solution is either to be a judge or appoint a jury, preferably as unbiased as possible. Get a devil's advocate too. You can rotate these people and pick some who do not know much about the person mentioned to be impartial and prevent oligarchies from forming. Look at number of mistakes and their severity, flaws. Check when and if there are successes and the person working above and beyond the standard of service or overcoming adversity. Look for initiative and improvements. It works almost exactly like military promotion if you think about it. It's a very human definition and not one turned into a metric. It's what we use and should use in courts of law too. There may be some guidelines to help judging but they should not be hard law as that way lies bureaucracy and lack of adaptability. Ultimately, all metrics depend on someone's judgement so why not make it explicit and transparent. The trickiest part is to maintain fairness. Fairness as defined by the subject accepting the judgement, which is different from equality. This generally comes with informing them of why a certain judgement has been made which is contrary to what happens in most companies - both in hiring and advancement.

AstralStorm 8y ago

Trying to turn running a software project or a company into an algorithm is a definition of bureaucracy. You do not look at metrics but judge performance. The solution is either to be a judge or appoint a jury, preferably as unbiased as possible. Get a devil's advocate too. You can rotate these people and pick some who do not know much about the person mentioned to be impartial and prevent oligarchies from forming. Look at number of mistakes and their severity, flaws. Check when and if there are successes and the person working above and beyond the standard of service or overcoming adversity. Look for initiative and improvements. It works almost exactly like military promotion if you think about it. It's a very human definition and not one turned into a metric. It's what we use and should use in courts of law too. There may be some guidelines to help judging but they should not be hard law as that way lies bureaucracy and lack of adaptability. Ultimately, all metrics depend on someone's judgement so why not make it explicit and transparent. The trickiest part is to maintain fairness. Fairness as defined by the subject accepting the judgement, which is different from equality. This generally comes with informing them of why a certain judgement has been made which is contrary to what happens in most companies - both in hiring and advancement.