3 ms·

The ZDNet article links to the HBR article [0]. Which traces to this networking measurement article [1]. [0] https://hbr.org/2022/06/why-microsoft-measures-em

The ZDNet article links to the HBR article [0]. Which traces to this networking measurement article [1]. [0] https://hbr.org/2022/06/why-microsoft-measures-employee-thriving-not-engagement?utm_source=pocket_mylist https://hbr.org/2022/06/why-microsoft-measures-employee-thri... [1] https://hbr.org/2020/08/can-you-be-too-well-connected https://hbr.org/2020/08/can-you-be-too-well-connected

Groxx 4y ago
Both of those mention "network" without really saying what they consider a network, and without even an average number, as far as can tell :/ they spend much more time talking about how many meetings and emails are involved. (And their meeting load sounds low! Dang, all of my managers have had like 30+ per week. Though I don't think I've ever sent as many emails as even their low end lists. Maybe they're just avoiding Teams...)
- caminante 4y ago
  > and without even an average number, as far as can tell Taking their analysis at face value, the average/median doesn't matter. This was a relative comparison for people that were (i) motivated; and (ii) had work-life balance. "network" was a function of how many people you interacted with via email/meetings. It's not surprising that people with better work life balance had less interactions. It is slightly surprising (IMHO) that people with less interactions were also more motivated.
  - Groxx 4y ago
    It matters how much of a reduction 17 implies. Scale-free numbers can be extremely misleading. E.g. last half, I had about 350 separate people I DM'd. Per year I generally break 500. If you include the size of email groups I have sent messages to, I'm consistently over 1,000 most months, possibly over 2,000 (maybe even significantly more). Microsoft has around 180,000 employees, that number could trivially go much much higher if I worked there. 17 is nothing if that's the criteria. Probably just noise, and you'll likely get different results next year. For people I would consider to be my network... 30-40? Maybe 50? They're the ones I actually interact with repeatedly on purpose. Reducing that by 17 is substantial, and I don't think I'd be as useful if I cut that size.
    - caminante 4y ago
      > Microsoft has around 180,000 employees, that number could trivially go much much higher if I worked there. > Reducing that by 17 is substantial, and I don't think I'd be as useful if I cut that size. IMHO, you're missing the point. They're allegedly controlling for the work environment by sampling and comparing peer groups at the same firm. Regardless of whether you think 17 is the right number for your situation, it's remarkable that there's a dividing line of 17 between the two peer groups in the study. Note that the 17 doesn't necessarily preclude someone having an even bigger network size delta from the non-thriving/non-balanced employees at MSFT.
      - Groxx 4y ago
        Edit: maybe a better way to phrase this is that, without knowing criteria and scale, there's really no way to know if this is surprising and possibly informative, or utter nonsense and it's hard to find anything conclusive because they're trying to hide their horrifying abuse of statistics. I see the latter happen far more often than the former, which is why significant claims require significant evidence. So far this seems weirdly vague. --- What I mean is that, depending on the criteria, their average network size could be 30 or 50,000. A reduction of 17 from 50,000 is only 0.03%. That's overwhelmingly likely to be statistical noise and nothing else. But given enough dimensions to slice the data across, that could be the single strongest correlation, so it's the one they claim. You see that kind of thing all the time in bad data science. The claim could fall anywhere between "shockingly strong correlation" and "utterly meaningless". So far I haven't been able to find anything that implies one way or the other.

Groxx 4y ago

Both of those mention "network" without really saying what they consider a network, and without even an average number, as far as can tell :/ they spend much more time talking about how many meetings and emails are involved. (And their meeting load sounds low! Dang, all of my managers have had like 30+ per week. Though I don't think I've ever sent as many emails as even their low end lists. Maybe they're just avoiding Teams...)

caminante 4y ago
> and without even an average number, as far as can tell Taking their analysis at face value, the average/median doesn't matter. This was a relative comparison for people that were (i) motivated; and (ii) had work-life balance. "network" was a function of how many people you interacted with via email/meetings. It's not surprising that people with better work life balance had less interactions. It is slightly surprising (IMHO) that people with less interactions were also more motivated.
- Groxx 4y ago
  It matters how much of a reduction 17 implies. Scale-free numbers can be extremely misleading. E.g. last half, I had about 350 separate people I DM'd. Per year I generally break 500. If you include the size of email groups I have sent messages to, I'm consistently over 1,000 most months, possibly over 2,000 (maybe even significantly more). Microsoft has around 180,000 employees, that number could trivially go much much higher if I worked there. 17 is nothing if that's the criteria. Probably just noise, and you'll likely get different results next year. For people I would consider to be my network... 30-40? Maybe 50? They're the ones I actually interact with repeatedly on purpose. Reducing that by 17 is substantial, and I don't think I'd be as useful if I cut that size.
  - caminante 4y ago
    > Microsoft has around 180,000 employees, that number could trivially go much much higher if I worked there. > Reducing that by 17 is substantial, and I don't think I'd be as useful if I cut that size. IMHO, you're missing the point. They're allegedly controlling for the work environment by sampling and comparing peer groups at the same firm. Regardless of whether you think 17 is the right number for your situation, it's remarkable that there's a dividing line of 17 between the two peer groups in the study. Note that the 17 doesn't necessarily preclude someone having an even bigger network size delta from the non-thriving/non-balanced employees at MSFT.
    - Groxx 4y ago
      Edit: maybe a better way to phrase this is that, without knowing criteria and scale, there's really no way to know if this is surprising and possibly informative, or utter nonsense and it's hard to find anything conclusive because they're trying to hide their horrifying abuse of statistics. I see the latter happen far more often than the former, which is why significant claims require significant evidence. So far this seems weirdly vague. --- What I mean is that, depending on the criteria, their average network size could be 30 or 50,000. A reduction of 17 from 50,000 is only 0.03%. That's overwhelmingly likely to be statistical noise and nothing else. But given enough dimensions to slice the data across, that could be the single strongest correlation, so it's the one they claim. You see that kind of thing all the time in bad data science. The claim could fall anywhere between "shockingly strong correlation" and "utterly meaningless". So far I haven't been able to find anything that implies one way or the other.