Slashdot Mirror


Former MySQL CEO Mårten Mickos Talks About Managing Remote Workers (Video)

Millions of pixels have been used to talk about Yahoo! CEO Marissa Mayer's decision to ban telecommuting and her reasons for doing it. Today's interviewee, Mårten Mickos, built MySQL AB into a billion-dollar company with 70% of its workers, all over the world, telecommuting instead of working in offices. Now he's CEO of another young open source company, Eucalyptus, and is following a similar hiring pattern. Mårten says (toward the end of the video/transcript) that he believes people working out of their homes is entirely natural; that this is how things were done for thousands of years before the industrial revolution.

35 of 100 comments (clear)

  1. Performance Metrics by djl4570 · · Score: 4, Insightful

    All you need is a method to accurately measure productivity.

    1. Re:Performance Metrics by Aviancer · · Score: 5, Insightful

      Pfft. Everyone on any given team knows who is good and who is dead weight. Listen to people, and make appropriate decisions. Yes, metrics are good to show improvement over time, but a weak, immature and cowardly way to identify poor performers.

    2. Re:Performance Metrics by h4rr4r · · Score: 2

      Which is exactly the same thing you need with in office workers. It changes nothing.

    3. Re:Performance Metrics by Kenja · · Score: 4, Insightful

      How about "do they do their job". This metric can be applied to everyone, not just those telecommuting. Now I work from home, writing software for very dull business processes. I do not think I would be nearly as productive at a desk 9-5 simply because some of my best ideas and breakthroughs happen at night when I'm winding down.

      --

      "Have you ever thought about just turning off the TV, sitting down with your kids, and hitting them?"
    4. Re:Performance Metrics by firex726 · · Score: 4, Insightful

      And therein lies the problem with what Yahoo was doing.

      They had few performance metric and feel that by having the seats warm they will get more productivity.

      So they just moved the slackers from the home to the office.

    5. Re:Performance Metrics by Anonymous Coward · · Score: 3, Insightful

      unless the recently dismissed was able to game your metrics to show they were the highest performing member of your team, and uses that data against you to win their case.

    6. Re:Performance Metrics by gweihir · · Score: 5, Insightful

      Indeed. And metrics are easily gamed or done wrong. The results are usually strategic mistakes, sometimes severe, as Yahoo will likely find out.

      There is no replacement for a competent manager with high personal integrity that actually has a well-founded expert opinion about all of the ones he manages. I see a primary task of a manger to create the environment where those in his/her care can work with the least amount of trouble from outside and at their most productive. This means that a manager actually serves his/her "underlings", not the other way round. Every good manager understands that.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    7. Re:Performance Metrics by gweihir · · Score: 4, Interesting

      And therein lies the problem with what Yahoo was doing.

      They had few performance metric and feel that by having the seats warm they will get more productivity.

      So they just moved the slackers from the home to the office.

      Indeed. This will not help. On the other side, they will severely inconvenience and anger a lot of those that worked productively from home. And these can easily get other and (now) better jobs. All this move does is to consolidate the low performers and make them the core of the workforce of Yahoo.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    8. Re:Performance Metrics by Anonymous Coward · · Score: 2, Insightful

      All you need is a method to accurately measure productivity.

      I really hate it when people just casually blurt that out as if it solves everything. Usually its an MBA or a manager saying that too. I am a QE team lead for two different small QE teams. I personally have an incredibly hard time to come up with any objective way to measure my personal productivity, especially on day-to-day granularity. Even for my testers, how am I supposed to objectively measure their productivity? By how many issues they test and close off? As ludicrous as measuring a developer on lines of code. By how many bugs are caught by them? Thats a measure of how bad the code is, not how good QE is. By how many issues are logged afterwards by customers? Even if thats a valid measurement (and its more a measurement of the complexity of features and how many customers might use a particular feature) - how does that translate to a day-to-day measurement? Or even to a measurement for an *individual*?

      'Metrics for productivity' is, in many cases, a phrase management loves because they can see some number go up or down, but in the real world, in complex environments, there simply is no simple way to 'accurately measure productivity'. There are few people just cranking out identical widgets every day and you just count how many widgets were produced.

      So I have to use my own judgement to see if my team members are pulling their weight and contributing.

    9. Re:Performance Metrics by mcmonkey · · Score: 4, Insightful

      There is no replacement for a competent manager with high personal integrity that actually has a well-founded expert opinion about all of the ones he manages.

      Right on. If you don't trust your managers, or don't know which managers to trust, you've already lost and all the metrics in the world won't help.

      To me this is the same issue as standardized tests. If you don't (or can't) trust your teachers, testing won't change that feeling. But how does the governor of a state know which teachers can't be trusted or should be replaced? She doesn't and shouldn't need to.

      The teacher in the classroom identifies which students are falling behind and need more help. The department head gets summary reports on student performance and monitors teachers. The school head gets summary reports on teacher performance and monitors the department heads. The head of the district gets summary reports on department performance and manages the school heads. And so one up to the governor, president, etc.

      No one other than the teacher in the room and that student's parents should be involved with an individual student's day-to-day performance. Not that the department head doesn't care about students, but the best way to express that concern is by putting the best teachers in place and giving those teachers the resources they need.

      Likewise, if the CEO is concerning herself with the day-to-day productivity of individual contributors, sounds like a company with one foot in the grave and the other on a banana peel. The CEO should be able to pick trust-worthy executives. Those executives should be able to pick able department heads. Department heads should oversee managers. Managers should manage people.

      That there may be a few slackers here and there is not a moral failing. But for the issues in a company to be so wide spread as to require a policy change of the magnitude we're seeing at Yahoo, you've got bad managers no able to motivate or replace bad workers, and bad department heads not able to identify bad managers, and bad executives not able to identify bad department heads, and a bad CEO not able to identify bad executives.

      Now that may be the case with Yahoo, which is why there is a new CEO and that CEO is making these changes. But she is bound to fail. The CEO should be concerned with getting the right executives in place. Those executives can retrain/replace department heads as needed. The department heads can get the managers on the right track. And then those managers can decide who needs to be in the office and who can work best remotely.

      A CEO jumping over a half dozen (or more) layers of management to tell a worker how to do their job makes as much sense as having the Secretary of Education sit down with each individual 3rd grade to check their sums.

    10. Re:Performance Metrics by Synerg1y · · Score: 2

      In a properly managed system from the start, I would completely agree, but once you start slipping, sometimes the only way is to "restart", go back to the traditional model and expand from there.

      My comparison is Jeans in the workplace, they're preferred by most employees and a lot of places just don't have a business need for business casual so employees wear Jeans, but when employees start showing up in faded Jeans with chains & logos, they're going to lose that privilege because at this point their dress code is interfering with the business. Same with remote workers, if they're managed properly and have clear expectations, no problem, you can actually save money through reclaimed office space, but if you have people on the payroll remotely, and you don't know what they do, that's when decisions like TFA are made. Let's be honest here, it's not like Yahoo's star was shining brightly before this change. It'll be interesting to see the results a year or two in the future.

    11. Re:Performance Metrics by Bucc5062 · · Score: 3, Insightful

      That comparison would only seek to enforce what the OP stated. If employees have a dress code, whether it is Business casual or jeans then it is up toe the immediate management to enforce that policy. If it gets/got out of hand then it is the immediate manager's problem, not hte company. To change that example, I come to work every in my clean, un-holed jeans wearing a decent shirt and shoes. Around me folks attire is failing so my reward for others bad behavior is to be punished for following the existing rules and required to go back to wearing a suit, because a manager could not enforce a rule. That is just wrong thinking.

      The same then applies in telecommuting. There are ways to measure performance, be it completion times, lines of code or some other metric. If management fails to measure shame on them. If they measure and fail to act again shame on them. That is their job thus the term Management. take away my telecommute, because others are irresponsible is throwing the baby out with the bath water, a poor management practice and like the OP said, a company that is rotten in the core and will soon fall.

      --
      Life is a great ride, the vehicle doesn't matter
    12. Re:Performance Metrics by ShanghaiBill · · Score: 2

      Indeed. This will not help.

      Actually, it almost certainly will help. Some people are marginally more productive at home. More people are drastically less productive. So on average, moving everyone back to the office will be a win. People are beating up Yahoo about making people come to the office, but they are ignoring Google which has always been much more stringent about telecommuting.

      they will severely inconvenience and anger a lot of those that worked productively from home.

      The people most inconvenienced and angry will be those using their "day off" to skimp on daycare fees and catch up on Desperate Housewives.

      All this move does is to consolidate the low performers and make them the core of the workforce of Yahoo.

      You are assuming that the drop in performance when telecommuting is correlated with being an inherently low performer in the first place. I don't see any reason to assume that is true.

    13. Re:Performance Metrics by HornWumpus · · Score: 4, Insightful

      Remote workers have nothing to do with integration testing.

      What you say could just as easily happen in the office.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    14. Re:Performance Metrics by HornWumpus · · Score: 2

      You start integration testing earlier.

      In the office or out, if they aren't even looking at the alpha builds of one of their team they aren't managing, much less testing.

      Unless the modules are completely independent the team will have spotted the air thief. It might take longer when telecommuting, but incompetence speaks loudly.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    15. Re:Performance Metrics by gweihir · · Score: 2

      Actually, it almost certainly will help. Some people are marginally more productive at home. More people are drastically less productive.

      I would dispute that for knowledge workers that want to work from home. It does not make any sense. There are those that do not want to work from home, and they certainly should not be forced to. But knowledge work is all about putting oneself in the right frame of mind and for some, that is easiest at home. There is also the factor that you cannot get quite a few of the ones in the higher competence range to work for you if that means moving.

      The people most inconvenienced and angry will be those using their "day off" to skimp on daycare fees and catch up on Desperate Housewives.

      No, the ones most inconvenienced will be those that took the job because working at home was part of the deal and allowed them not to move to another city. Those that just slack off at home will be angry, true, but they are not an issue, I completely agree with you there. But they cannot leave easily either. Higher skilled people that only took the job because they would not have to move will likely not be too angry, but just move to a better employer. And that is the real problem.

      You are assuming that the drop in performance when telecommuting is correlated with being an inherently low performer in the first place.

      It is not obvious to you? I find that surprising. Keep in mind, I am only talking about people that want to work from home.

      From my experience, there are some that cannot focus at home, but then will seek out the workplace, i.e. they do not want to work at home anyways. I also saw this a lot in students: There are those that are low performers, no matter what. There are those that perform well, and some say they need a lab space to do it (we would always give them basically a regular workspace, including lockable table, etc.) and others that perform well, but say they work best at home. What I have never encountered was a high performer without a distinctive preference, or a low performer that was sabotaging him/herself by choosing the wrong work environment. That is among the stuff you figure out pretty fast, if given a choice. And one thing universally true for high-performers is that they want to be high-performers. They will try to fix anything that hinders them from being.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    16. Re:Performance Metrics by ShanghaiBill · · Score: 2

      Actually, it almost certainly will help. Some people are marginally more productive at home. More people are drastically less productive.

      I would dispute that for knowledge workers that want to work from home. It does not make any sense.

      It may not make sense, but that is the way it is. My company allows telecommuting on a case-by-case basis, and for many people it does not work.

      But knowledge work is all about putting oneself in the right frame of mind and for some, that is easiest at home.

      Zoning out in front of the TV is also easiest at home.

      You are assuming that the drop in performance when telecommuting is correlated with being an inherently low performer in the first place.

      It is not obvious to you?

      No, it is not obvious, and it does not mesh with my experience. Some high-performers do better at home, some don't. Some not-so-high-performers do better from home, some don't. My experience is that the best predictor of how well telecommuting will work are:

      1. Does the employee have young children? Bad.
      2. Does the employee have a stay-at-home non-working spouse? Bad.
      3. Does the employee have a spouse that also telecommutes? Good: they keep each other honest.

      Of course, these are too politically incorrect to be actionable policy.

    17. Re:Performance Metrics by Zontar+The+Mindless · · Score: 2

      If I am *arbitrarily*, *for no reason related to my conduct* required to wear 3x as many clothes to do the same job, I'd tend to view that as a punishment, yes, and an unjust one at that.

      P.S. Home worker more often than not, sometimes in my underwear. In other news, water is still wet, and software manuals still get written. Film at 11.

      --
      Il n'y a pas de Planet B.
  2. one difference by Trepidity · · Score: 4, Insightful

    These companies did it on purpose and planned for it, while it sounds like it just sorta "happened" at Yahoo, with management neither having a plan for how to manage it nor (apparently) really paying any attention at all to what remote workers were doing and how they were doing it.

  3. Re:TFA by Anonymous Coward · · Score: 5, Informative

    Robin: Marten, what percentage of your MySQL workers work from home?

    Mårten: We had 70% working from home when we were 500 employees in total.

    Robin: Okay, okay. 70%?

    Mårten: We were based in 32 countries across 18 time zones.

    Robin: 32 countries, 18 time zones.

    Mårten: Yeah.

    Robin: How did you manage those workers?

    Mårten: I wonder if I did, meaning I mean something with it, I think when you manage a distributed team, you cannot manage through command and control; you must manage through vision and culture.

    Robin: Okay.

    Mårten: You must get the vision across to everybody. You must agree on how you behave, and what the company culture is. And then you let them do what they know they need to do. And that is how it works. But if you think you must observe them and monitor them and command them, and control them, then it won’t work for you.

    Robin: Okay. So you need very self-motivated people, that you are telling me.

    Mårten: Very true. I call it the fishing village analogy. Meaning our people at MySQL and now at Eucalyptus are like fishermen. They live in a fishing village and are very social together, but every morning before the sun dawns, they go out in their small boats to sea and they are all on their own, and they come back only when they have caught fish.

    Robin: Okay, now recruiting these fishing people, (that’s a beautiful analogy) recruiting these independent workers, is it different from recruiting people you are going to be able to watch at their desks?

    Mårten: Yes and no. First you have to interview them like you do with anybody, you have to post your open reqs like you do with anybody, but of course you must check that they truly belong to the portion of the world population that is capable of working from home, because not everybody is. It is not for everybody. It is for some of the best people in the industry but it is not for everybody.

    Robin: Okay, and in the industry, what jobs work best filled by remote workers? And what works worst?

    Mårten: As main rules I would say if your product is an intangible product, then it works well. And it so happens that software was the first industry to do it, but you can do it in politics, medicine, science and arts as well. The second rule is that for this to work, people need to be able to go all in online. They need to be able to live not just their professional life, but convey their personality online as well. Because the argument against distributed teams is that body language doesn’t work, and you don’t get the sort of the closeness, but on the contrary, we say no, that is not true. You can bring your personality and even your body language online if you decide to do so. And that is how you make it work.

    Robin: Now you are talking about creative people, programmers, scientists, the artists; what about people like finance and marketing? Are they good, as good remotely?

    Mårten: They are. And I would say their job is increasingly creative. But we had people working from home in every part of the company. We had accounts receivables, which was operated as a home operation, marketing was done, some of the accounting as well; of course, there are functions where you have to be in an office, you have to put things on real paper and store them in a real cabinet. So I am not saying you can live completely without it. But I don’t see any part of the organization that couldn’t be at least partly distributed among people who work from their homes.

    Robin: How much money does it save if you have a quantification, how much does it save with all these people working from home?

    Mårten: We never did it for the purpose of saving money. And we told ourselves that what we saved in office costs, we spent in travel costs. And that is probably more or less true. Maybe we saved a little bit but not much. A benefit we got and whether that is a saving or not I don’t know, but we managed

  4. Engaging work by s1d3track3D · · Score: 4, Insightful

    If you have a cool product, interesting things to do and hire interested people, you will have good employees.
    Many technical people work in the field because they enjoy it, how many people work on FOSS in their spare time anyway?

    Working on new, interesting, challenging things is fun! Maintaining 'legacy' stuff, not as fun. No disrespect to Yahoo but Flicker, Yahoo Mail, YUI, OMG! (please), for me it would be hard to be excited about maintaining these.
    Additionally, working in a smaller company where one person can really help shape things is huge, being just another worker bee in a huge corporate environment can be depressing. (especially one with a declining public image)

    Obviously, just my opinion.

  5. Re:The industrial revolution improved productivity by Daetrin · · Score: 4, Funny

    Not a nasty, dirty, wet cave, filled with the ends of worms and an oozy smell, nor yet a dry, bare, sandy cave with nothing in it to sit down on or to eat: it was a hobbit-cave, and that means comfort.

    --
    This Space Intentionally Left Blank
  6. Re:MySQL built for it by TankSpanker04 · · Score: 2

    Well put. I would add that working remotely works well for some companies and not others -- regardless of the industry. One might think Yahoo! being a tech giant would automatically mean remote-friendly. But as you said, it depends on culture and planning. I'm willing to bet Ms. Mayer did her homework before coming to this conclusion. Perhaps the majority of their remote workers really were either not pulling their weight or were becoming hermits.

  7. Really hard to do. by Chirs · · Score: 4, Insightful

    What do you use for measuring "productivity"?

    Lines of code? My happiest work days are when I end up removing more code than I put in. Also, this is really easy to game.

    Bugs fixed? I usually end up working on the really nasty bugs...intermittent, only occur in customer sites, and under no circumstances can you shut down the system to debug it. Some bugs take weeks or months to track down.

    Hours worked? Pointless, doesn't track if you're actually being useful during those hours.

    While it's easy to measure productivity if you're making widgets, its *really hard* to measure productivity if you're doing creative stuff.

  8. how do you measure? by Chirs · · Score: 2

    What metrics will tell you that someone is doing good work?

    Suppose I mostly review other people's code and make suggestions for improvement, answer lots of random questions about obscure corners of various specifications, work on really tricky bugs that take a long time to track down, look at upcoming roadmaps and figure out how they're going to affect us, etc.. What objective metric do you use to measure my performance? Lines of code submitted doesn't work, bug closure rate doesn't work--there is no simple numerical statistic to measure.

    1. Re:how do you measure? by RabidReindeer · · Score: 3, Insightful

      Lies, Damn Lies and Metrics. Metrics are what you use so that you can be blind to things that can't be/aren't measured.

      Lines of Code is a classic. If I rip down an algorithm and replace it with one that's faster, more reliable, and 1/3 the size, I have negative LOC. And if LOC is all of the above that got metrics, I'm a loser.

      I have a great respect for being able to measure things (where it's possible and meaningful), but unlike the legendary Statistical Bikini, it's what isn't covered that's as important as what is. Metrics should be a guide, but when you have people spending all their time fiddling with metrics, and other people spending time fiddly ways to look good under the metrics you're losing a lot of your productivity to metrics. And unless you are in the business of producing metrics, that's not good.

    2. Re:how do you measure? by ShanghaiBill · · Score: 2

      What objective metric do you use to measure my performance?

      Peer stack ranking. If your peers think you are deadweight, they you probably are. You can brown-nose or snow-job one manager, but you can't fool a whole team.

    3. Re:how do you measure? by HornWumpus · · Score: 3, Insightful

      Works while you have functional teams. When the team is dis-functional, not so much. Then it becomes a popularity contest.

      --
      John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
    4. Re:how do you measure? by seebs · · Score: 2

      So, you're saying, the guy who solves the hardest problems that no one else could reasonably attempt is the worst employee? Because that guy's going to take more time to solve things, is gonna have more things that need to be reworked, and so on.

      Think it through! If you have a simple task that any reasonably competent engineer could do, an average programmer should be able to do it quickly and reliably, right? So then you have stuff that's basically new research and trying to solve problems that no one's yet sure how to solve, and...

      1. We won't have any way to make reasonable estimates, so estimated dates will be missed sometimes.
      2. Amount of output will be small.
      3. Likelihood of rework is high.

      Look, seriously, if it were that easy, people would already be doing it widely enough that there wouldn't be a need to invent metrics, we'd have working ones. We don't.

      --
      My blog: http://www.seebs.net/log/ --- My iPhone/iPad app: http://www.seebs.net/seebsfrac/
    5. Re:how do you measure? by RabidReindeer · · Score: 2

      Better stated: you need to know the context in which the metrics are stated.

      That's only a small part of the problem. Factories in the 1960s were quite aware that they were spewing sulfuric acid. Their metrics failed to consider just how outraged people could become. You can fail because you didn't measure something critical or you can fail because you gave the wrong weight to something you did measure. And you can fail because you attempted to extrapolate a metric using the wrong functions.

      If we ever actually managed to obtain truly accurate and meaningful metrics that actually mapped reality accurately, management would be something that you could replace with software. But just as software design has consistently failed to be something that can be reduced to a mindless rote process, so too has good decision-making.

      Metrics are no more a "silver bullet" than 4GLs, OOP, CORBA, AJAX or countless other one-size-fits-all solutions. They are a tool. When they become disproportionately valued, the management process suffers.

  9. A job-related article with no link to Dice! by olip85 · · Score: 2

    Are things going back to normal?

  10. Worse than that. by khasim · · Score: 2

    They had few performance metric and feel that by having the seats warm they will get more productivity./blockquote>From the other article, the "metric" was that the new CEO thought that the parking lot was too empty.

    So it isn't even "performance" but more like "attendance".

  11. Home Work was Piece Work by westlake · · Score: 2

    he believes people working out of their homes is entirely natural; that this is how things were done for thousands of years before the industrial revolution.

    Divide and conquer.

    If you were producing for a market, home work was piece work --- with no labor laws or labor unions to prevent abuses.

    When the textile mills of New England began opening jobs to young women --- their first taste of independence, education, organization and a real, substantial, pay check --- girls abandoned the rustic life and never looked back.

    In union they found strength.

  12. Re:My 0.02 by Darinbob · · Score: 2

    For some employees, they will be productive at home. Maybe even more productive than at the office. But for some different employees it will be worse. People actually do slack off at home, they will try to get only the minimum done, they will try to game the system. It is only a minority of employees who are self starters who are willing to work as hard as they can without self supervision as long as someone waves a stock option on a stick. But I see a lot of companies who mistakenly assume that most employees are that way and that employees enjoy working as hard as they can. The companies are probably naive in assuming employees are actually loyal.

  13. Re:Everyone on any given team knows who is good by ILongForDarkness · · Score: 3, Insightful

    Also a lot can be team dynamics. A person can be a great worker but for whatever reason a few other team members decided that they don't like talking to them. So they are never included in conversations aren't seen as helpful when problems come up etc. But is it due to a real personality fault in that employee or that employee just having a different way of communicating, work style heck even extra curricular interests can come into play (people will generally go to the person that they can chat with for a half hour about the latest sports drama than the guy that is say a dungeon master (when sports are their interest and not role playing) or vis versa). That is part of the issue with remote work that needs to be considered not just individual work performance but how well will the team communicate without the queues you get from in person interaction? It can work and it can not work but you need to at least leave the option of going back to a work from the office model if the telecommute doesn't work for the employee (or you find other people's performance goes down because they aren't as available for helping out with random questions etc).