5 ms·

This should be a monthly census: it needs to be automatic. If you're the manager of a company and you are relying on people to say something about their co-work

This should be a monthly census: it needs to be automatic. If you're the manager of a company and you are relying on people to say something about their co-workers, then you will miss most of the information that you could be capturing, and you will get highly biased impression of your workforce, because only a certain type of person will come forward to praise their co-worker. There are various voting systems that work well, but the easiest is something like "Every month, list 3 of your co-workers who did something impressive." This avoids the many problems with something like the stack-rank system that Microsoft was committed to, for too many years. In that system, workers were forced to list their co-workers, not just as "good", but also a certain percentage had to belong to the "bad" group, the lowest 30%. Most workers do not want to attack their fellow workers this way. You can get all the information you need by simply letting your workers upvote each other. You should ration the upvotes so that they are meaningful. On a small team, perhaps people get 1 or 2 votes, in a big company, perhaps they got 5 or 6 votes. If you make the votes a scarce resource, you get better accuracy: people think carefully about who they should upvote. There will be some workers who never get upvoted by anyone. These are bad workers. There will also be some workers who get very few votes, or who only get votes from "rings" they've formed with other workers, where they vote for each other. These are suspect workers. In all cases you'll want to apply something like the Google PageRank algorithm to the upvotes. That is, an upvote from someone who is themself highly upvoted should matter more than an upvote from someone who rarely gets upvoted. I would not ban people from the article's main idea "When your coworker does great work, tell their manager" however, keep in mind, if you are the manager or CTO or CEO, you need to ensure there is also a system in place that is more automatic than simply waiting for people to talk about their co-workers.

loopz 6y ago
Who are you to know who is doing good work or not? If managers are not involved, they should have no say.
- lkrubner 6y ago
  People do reliably know which of their co-workers are good and which of their co-workers suck badly. But also, the rule of 150 applies here. At a startup with less than 100 people, each worker often has an idea of everyone's strength, throughout the company. However, if you are at a big company, with thousands of employees in multiple locations, then it is often best to restrict upvoting to some subset of the company. For instance, a company with 1,000 workers might have 10 teams of 100 people each. At such a company, it might make sense to have the rule that you can only upvote people who are on your team. The assumption is that these would be the people you interact with and have a good idea about who is good and who is bad. Having said all this, I'm not suggesting that this should be the only way that managers evaluate their workers, but having a formal, structured, regular process is important. There are other techniques that a good manager should also follow. If you have time, you can listen to the interview I gave here: http://simpleleadership.io/why-group-meetings-can-be-time-wasters-with-lawrence-krubner/ http://simpleleadership.io/why-group-meetings-can-be-time-wa... Too many managers waste too much time in big meetings, more time should be spent in one to one meetings. About this: "Who are you to know who is doing good work or not?" I can't tell if you mean that as a criticism of my suggestion or whether you are attacking the original article for the idea "When your coworker does great work, tell their manager." Because, if you believe people can not evaluate their co-workers, then you must disagree with the above article.
  - loopz 6y ago
    Every time this has been attempted, it devolves into toxic work culture. Also, how can you know what other people are working on? Most of the work is outside anyone's scope, and end up toxic when optimizing for arbitrary metrics. I speak for myself, mostly unrelated to articles put on www. I don't have to agree or disagree by anyone else's metrics and dominating rhetorics. Not meant to disrespect, but am not fully aboard naive beliefs. Experience and learning from others will in time tell, and moderate good intentions.
    - lkrubner 6y ago
      No, I think you are thinking of systems like the stacked rank system they used at Microsoft.
T-hawk 6y ago
Wouldn't co-worker upvotes lead to workers prioritizing their output for what attracts upvotes rather than what the company needs to get done? You'd get a team constantly making tools for each other and no sellable product.
- ping_pong 6y ago
  Yes, this definitely happens. The incentive system should align with good behavior, otherwise you get massive gaming of the performance system. I've seen this in a few companies I've been at, moreso the large ones because no one can keep complete track of everyone's gaming of the system.
- csharptwdec19 6y ago
  Yes. It's almost a reflection of 'Golden Parachute CEO' behavior: There's an implicit benefit to making decisions that have 'good optics' at the moment even if it takes you down a road you'll regret. Seen this a LOT with 'buy/build' decisions almost everywhere I've worked. The cost savings look great up front, but by the time you've added all your custom bits, you're hitting things like API limits (either in surface or billing) and having to shell out, or building so much boilerplate to get around the limits you may as well have built it yourself. But hey, it saved X$ the first year, who cares if it costs more the next 5 to maintain?
- lkrubner 6y ago
  In terms of gaming the system, that is largely avoided by use of the Google PageRank algorithm, as I mentioned below. On a team of 100 you don't have to deal with the millions of spammers that exist on the Internet. On a team of 100, the Google PageRank algorithm works almost perfectly. But as I also mentioned, any 360 tool should be just one tool that a manager uses, it should not replace all other tools. I'm curious what you thought of the original article and its idea "When your coworker does great work, tell their manager". Are you opposed to that? If you are not opposed to that, then why would you be opposed to any other system of feedback, especially those that are formalized and structured?

loopz 6y ago

Who are you to know who is doing good work or not? If managers are not involved, they should have no say.

lkrubner 6y ago
People do reliably know which of their co-workers are good and which of their co-workers suck badly. But also, the rule of 150 applies here. At a startup with less than 100 people, each worker often has an idea of everyone's strength, throughout the company. However, if you are at a big company, with thousands of employees in multiple locations, then it is often best to restrict upvoting to some subset of the company. For instance, a company with 1,000 workers might have 10 teams of 100 people each. At such a company, it might make sense to have the rule that you can only upvote people who are on your team. The assumption is that these would be the people you interact with and have a good idea about who is good and who is bad. Having said all this, I'm not suggesting that this should be the only way that managers evaluate their workers, but having a formal, structured, regular process is important. There are other techniques that a good manager should also follow. If you have time, you can listen to the interview I gave here: http://simpleleadership.io/why-group-meetings-can-be-time-wasters-with-lawrence-krubner/ http://simpleleadership.io/why-group-meetings-can-be-time-wa... Too many managers waste too much time in big meetings, more time should be spent in one to one meetings. About this: "Who are you to know who is doing good work or not?" I can't tell if you mean that as a criticism of my suggestion or whether you are attacking the original article for the idea "When your coworker does great work, tell their manager." Because, if you believe people can not evaluate their co-workers, then you must disagree with the above article.
- loopz 6y ago
  Every time this has been attempted, it devolves into toxic work culture. Also, how can you know what other people are working on? Most of the work is outside anyone's scope, and end up toxic when optimizing for arbitrary metrics. I speak for myself, mostly unrelated to articles put on www. I don't have to agree or disagree by anyone else's metrics and dominating rhetorics. Not meant to disrespect, but am not fully aboard naive beliefs. Experience and learning from others will in time tell, and moderate good intentions.
  - lkrubner 6y ago
    No, I think you are thinking of systems like the stacked rank system they used at Microsoft.

T-hawk 6y ago

Wouldn't co-worker upvotes lead to workers prioritizing their output for what attracts upvotes rather than what the company needs to get done? You'd get a team constantly making tools for each other and no sellable product.

ping_pong 6y ago
Yes, this definitely happens. The incentive system should align with good behavior, otherwise you get massive gaming of the performance system. I've seen this in a few companies I've been at, moreso the large ones because no one can keep complete track of everyone's gaming of the system.
csharptwdec19 6y ago
Yes. It's almost a reflection of 'Golden Parachute CEO' behavior: There's an implicit benefit to making decisions that have 'good optics' at the moment even if it takes you down a road you'll regret. Seen this a LOT with 'buy/build' decisions almost everywhere I've worked. The cost savings look great up front, but by the time you've added all your custom bits, you're hitting things like API limits (either in surface or billing) and having to shell out, or building so much boilerplate to get around the limits you may as well have built it yourself. But hey, it saved X$ the first year, who cares if it costs more the next 5 to maintain?
lkrubner 6y ago
In terms of gaming the system, that is largely avoided by use of the Google PageRank algorithm, as I mentioned below. On a team of 100 you don't have to deal with the millions of spammers that exist on the Internet. On a team of 100, the Google PageRank algorithm works almost perfectly. But as I also mentioned, any 360 tool should be just one tool that a manager uses, it should not replace all other tools. I'm curious what you thought of the original article and its idea "When your coworker does great work, tell their manager". Are you opposed to that? If you are not opposed to that, then why would you be opposed to any other system of feedback, especially those that are formalized and structured?