So, apparently Google found that of their 8 characteristics-of-good-bosses, technical excellence was the least importance. Fair enough. But note that this is looking at a sample of technically excellent bosses! So, e.g., Bock is quoted as saying: "In the Google context, we’d always believed that to be a manager, particularly on the engineering side, you need to be as deep or deeper a technical expert than the people who work for you."
It doesn't at all follow that technical excellence isn't extremely important for bosses who haven't been chosen for technical excellence. (Perhaps it isn't; but Bock's work doesn't seem to provide any evidence.)
The article actually said that technical excellence was important, it just said that it wasn't as important as some of the people skills.
I can't really see Google dropping technical competence as a qualification for managers. I've heard that from 2005-2007 (before my time), Google brought on a bunch of non-technical managers from other companies, and the result was a pretty uniform disaster. Most of those managers are no longer with the company.
Reverse-engineering from the article, the original statistical result was probably something like, "Variations among managers in technical skill account for very little of the variations among managers in performance reviews."
One way to get that result would be for technical skill to be unimportant. Another way would be for the variations in technical skill within the sample to be much smaller than other relevant variations. Given that (según the article) the managers were selected for technical skill, but not for the other qualities, it seems almost certain that this was the case.
This might sound like a bug that could be avoided by better statistical analysis techniques, but it's really not. You really want to know what causes the actual variations in performance in your sample, not what could conceivably cause variations in performance. Surely there are hundreds of things that would cause even bigger variations than what's in your actual sample — you could have managers who kill all their employees, or managers who are interned incommunicado in Guantanamo, or managers who obsessively turn every project into an investigation of fluoride contamination of their precious bodily fluids, or managers who are actually undercover headhunters from a competitor trying to steal away all of your best talent.
Cosma Shalizi has an excellent post about the pitfalls of multivariate statistical analysis at http://www.cscs.umich.edu/~crshalizi/weblog/520.html; although he's talking about genetics, it's very likely that what he's saying applies to Project Oxygen as well:
To see why gene-environment interactions matter, consider one of the best-established links between genetic variations and intelligence, phenylketonuria. This is a recessive genetic disease which interferes with the normal metabolism of the amino acid phenylalanine. If someone with one of the defective forms of the gene for phenylalanine hydroxylase consumes too much dietary phenylalanine, it leads, among other problems, to serious mental retardation. Under suitable diets low in phenylalanine, however, they grow up mentally normal. Assigning shares of this effect to the genes and to the environment is exactly as sensible as trying to say how much of the fact that a car can go is due to its having an engine and how much is due to their being fuel in the tank. The best the usual biometric model could do here would be to predict that having the gene always reduced intelligence, as did consuming phenylalanine (which would be bad news for makers of artificial sweeteners); the fact that it's the combination, and only the combination, which is a problem would be missed, and the predicted size of the effect would be badly wrong. … So while everyone piously says that genes and environments interact in development, they typically use models which assume that they do so only in trivial ways, and hope that any actual interactions are small enough to be treated as noise.
(So if these simple linear models are so bad, why does everybody use them? Because they have fewer parameters than more complicated models, which means that they're not as prone to overfitting. It's easy to construct a nonlinear multivariate statistical model with more parameters than you have managers in your company, which will discover correlations no more meaningful than the fact that people whose initials are "BHO" are far more likely to currently be the president of the United States than people with other sets of initials.)
Right, I'm saying that the managers in charge of interpreting this data are undoubtedly aware of this effect, and so they're looking for those effects besides technical skill. It's like every Google manager already has an engine (they took care of that problem several years ago), now what can they do to make sure that there's always gas in the tank?
I was disagreeing with your comment, "it just said that it wasn't as important as some of the people skills." The article did say that, but the statistics almost certainly did not. I agree with your new comment.
Microsoft learned this lesson a long time ago. In the old days of MSFT, the saying was that BillG believed that everyone's manager needed to be a better coder than him/her. The company learned eventually that the best engineers do not necessarily make the best managers and engineered the IC/manager split-career path that exists today.
P.S I like how Google is finding out they're just the same as pretty much any other large organization :).
Being there in the old days, I really liked two things.
1) Any question I had, someone could either give me the answer, direct me to the book (yeah, that long ago) that had the answer, or do the back and forth so I could figure out the answer with the knowledge I already had.
2) Rip the sh*t out of my code in a way that, while humiliating, made me a much better programmer.
You went in thinking that you're the smartest kid in the room, but while I was there, the mean IQ was 132. It was a place with a bunch of smart _adults_ who also had a significant amount of knowledge and experience, and very few hoarders of that information. After all, how can you show how smart you are if you keep quiet when someone is wrong? My niches were that I had K&R memorized (that's really embarrassing now, but "actually, free(NULL) is legal," beats a full house) and my proof of concept client code made its way into SDKs.
Joel Spolsky talks a lot about the same time era, basically Microsoft pre-Internet. Many of us ex-MSFTies idealize it to a certain degree, but that's fine–take and keep the good and drop the bad. Many of us are engineering leads/CTOs/flounders elsewhere, and the rigor present in the hierarchy of ability-to-execute helps to excel in other companies or industries.
But most of us secretly would like to go back to a few thousand nerds in a rainy forest and hack products all day. :-)
Mean IQ, so if we assume IQ is normally distributed and there're a few people with IQ > 170, that'd imply that the bulk of people have IQs in the 125-130 range. This seems eminently believable to me.
This was also the time when the legendary "Microsoft interview" started coming together. I.e., you'd have an all-day interview with everyone on your prospective team (if you were targeting a particular position), and each interviewer would have one or two IQ-test questions and at least one whiteboard coding exercise.
People really went after the lateral-thinking questions and the estimation questions to find "smart" people. "How would you calculate how much rain falls in Washington in a day,"[1] "How many gas stations are there in (IYFStateOrCity)," and "How would you make a desk calendar with two cubes,"[2] were all fun ones.
Now, by making your questions similar to those on IQ tests, you select people who tend to do well on IQ tests. Funny, that.
[1] I got a boost on this one by spitting out,"weigh the clouds on their way in and out of the state," right away before building a slightly more feasible estimate. :-)
[2] The trick is just like the Jimi Hendrix song, if six was nine.
Well, you also run into a practical testing ceiling. Most of the general purpose IQ tests (WAIS, Stanford-Binet) don't measure very accurately over 160. 132 as the mean of a set systemically limited to a maximum of 167 paints a different picture, not that I have a good handle on how many of us exceeded 160, especially across the entire campus.
Microsoft is still dramatically successful today. The problem is that people perceive differences in success, not absolute numbers. So if you were wildly successful ten years ago and you're still wildly successful, what the hell have you been doing? Why haven't you improved?
The problem Microsoft is facing is that they basically accomplished their mission statement. "A computer on every desk, all running Microsoft software." That is an accurate description of the world circa 2000. What the hell do they do now?
Google is likely to face a similar problem in the next few years. "Organize the world's information and make it universally accessible and useful." For all the flack they're getting lately, they've done a remarkably good job at it. People expect to have whatever information they need at their fingertips, and they get mad when they don't. You wouldn't see the vitriol toward bad search results that you do now if the Internet was like it was in 1998, because nobody really had an expectation that they'd be able to find what they were looking for then.
I think you may be underestimating the potential reach of Google's ambitions in the search space. Think of improvements in search in specific areas of knowledge. How about searching for flights? Well, they bought ITA software for that, probably because they wanted to get into that space. How about organizing books? Right, Google Books? Scientific papers, law information. The extent that searches and organizing information in various disciplines is virtually limitless. Then there are things like the semantic web. I would go as far as saying that almost any kind of problem solving humans do involves a form of 'search', and as these processes take advantage of the ever increasing information on the web, they will benefit from improved search techniques.
To conclude and restate again, Google's objectives, IMHO, are extremely far-reaching.
"Once they had some working theories, they figured out a system for interviewing managers to gather more data, and to look for evidence that supported their notions."
Did they also look for evidence that cast doubt on their notions? Looking for both kinds of evidence, and evaluating them fairly, would lead to a more robust conclusion.
I don't want to disparage your comment because I do agree to some extent.
However, I think you might be getting voted up because you gave links to some interesting content instead of because people think you are necessarily correct.
I'd like to provide a little balance.
It's routine practice in statistical analysis of situations to form a hypothesis which goes against the grain of accepted belief and then to test it using analytical techniques [1]. In fact, this is how the scientific process works.
Given that we're discussing one of the most sophisticated data analysis companies around they probably did attempt to test their theories to destruction. Or, otherwise stated, after forming an initial hypothesis they tried different interpretations of the same data to see which ones made distinguishable testable propositions.
And yet, when its phrased like that, it (almost?) seems perfectly reasonable. It's easy to miss the "obvious" issue, and I've been following the skeptic scene for years!
>One of the brighter students (by the name of L. da Vinci) was immediately promoted to the manager of the project, putting him in charge of procuring paints, canvases, and brushes for the rest of the organization.
Everyone, except businesses (it seems), knows that technical skill != management skill, and the best workers will not necessarily be even good managers. But such knowledge is so rarely followed.
If the guy outta Venice knew his stuff, he'd have immediately delegated those tasks to others and got back to his own business, rather than do them all himself.
I get to see companies of all sizes grapple with various problems and the most interesting thing I've noticed is that each company has a "culture" that influences everything it does. So if it's an insurance company, they're going to think in terms of risk and coverage. If it's a manufacturing company they're going to make decisions based on flow models and statistical process control. If it's an quasi-military organization, they're going to think in terms of hierarchical structures.
It's neat to see Google continuing this pattern by applying tagged data collection and statistical inference to their quest for organizational optimization.
I can only imagine what it's like to work in a large-scale IT operation in the porn industry ;)
"I can only imagine what it's like to work in a large-scale IT operation in the porn industry ;)"
I've heard it's about the most unsexy thing ever. The technical aspects are all about shipping massive quantities of data cheaply and reliably. The, erm, video-production aspects are about presenting a picture that will satisfy customers and leave them coming back for more.
In other words, it's pretty much like Google or any other big-data startup (and there're employees at Google who used to work in the porn industry). Not having been a part of it, I don't know what the culture is like, but I've heard it's pretty much all business, and they have significantly less fun than we do at Google.
So, typically there's a split between creative, project-based leadership and productive, people-based management. In games these creative leaders are (in the west) lead designers, Microsoft calls them "program managers," movies call them directors, and so on. Sometimes it's called "matrix management" and rigidly structured. Whereas at Blizzard, project management is just clearly set up so that the design staff are in complete creative control of the project's direction.
Anyway, I don't really see that separation of roles in evidence within the scope of this article. How does it work at Google? Also, is there some good analysis on this general subject someone can point me to?
Wow, they only just figured out that managers don't need to have a greater technical skill than those they are managing? I thought that was management 101?
Managing and engineering are 2 different skills. While managers benefit from technical knowledge it certainly doesn't need to be greater.
And with Marissa Mayers comments on Google wants to connect the real world to the digital world, where have they been. In case they weren't aware, internet connected appliances, location based apps, RFID, QR codes, mobile phones have been around for a little while now.
Google's search engine is great though slowly declining in quality but that is all they have been great at. Android is a decent mobile OS but it only received market pentration due to it's zero cost. Their social attempts have all failed. They only seem to get wide market adoption from free software or services, which I suppose they use to keep their search engine traffic high so they can maintain their advertising revenues.
They seem like a company unable to create another profitable software line.
Google’s approach optimizes a sub-optimal model. A poor performer is less bad after a year of coaching. Our data show that talent occurs in predictable combinations, each combo in predictable frequency. Some combos DO NOT exist at all. The talents of great leaders (envision future, inspire followers) does not exist in the same person as a great manager (sees unique talents in others, applies and develops the talent, fosters purposeful collaboration). Google should have asked: What talents do our best managers posses? How do we find and develop more, faster? See http://methodteaming.com/did-google-miss-the-forest-for-the-...
I notice they have no bullet points listed under "Help your employees with career development." I wonder if this means they feel obligated to mention it but don't really want to encourage it (because developing your career will often mean leaving) or if it just means they don't know how.
Or maybe it's supposed to be so obvious that it doesn't need explanation.
Probably the opposite, making sure that the employees can figure out what they want to do with their career, and know how to get there without leaving the company is a good retention tool.
It is very difficult to measure the quality of each manager on these eight effective habits in an objective manner.
What is the opinion of this group, about, google people rank algorithm(ratings from good performers carry higher weight and vice versa) to rate employees within the company.
The story links to the 8 behaviors in the form of a JPG image, but I transcribed it to put in the HR section of our wiki. Thought someone else might want the plaintext version:
EIGHT GOOD BEHAVIORS
1. Be a good coach
- Provide specific, constructive feedback, balancing the negative and the positive.
- Have regular one-on-ones, presenting solutions to problems tailored to your employee's specific strengths.
2. Empower your team and don't micromanage
- Balance giving freedom to your employees, while still being available for advice. Make "stretch" assignments to help the team tackle big problems.
3. Express interest in team member's success and personal well-being
- Get to know your employees as people, with lives outside of work.
- Make new members of your team feel welcome and help ease their transition.
4. Don't be a sissy: Be productive and results-oriented
- Focus on what employees want the team to achieve and how they can help achieve it.
- Help the team prioritize work and use seniority to remove roadblocks.
5. Be a good communicator and listen to your team
- Communication is two-way: you both listen and share information.
- Hold all-hands meetings and be straightforward about the messages and goals of the team. Help the team connect the dots.
- Encourage open dialogue and listen to the issues and concerns of your employees.
6. Help your employees with career development
7. Have a clear vision and strategy for the team
- Even in the midst of turmoil, keep the team focused on goals and strategy.
- Involve the team in setting and evolving the team's vision and making progress toward it.
8. Have key technical skills so you can help advise the team
- Roll up your sleeves and conduct work side by side with the team, when needed.
- Understand the specific challenges of the work.
THREE PITFALLS OF MANAGERS
1. Have trouble making a transition to the team
- Sometimes, fantastic individual contributors are promoted to managers without the necessary skills to lead people.
- People hired from outside the organization don't always understand the unique aspects of managing at Google.
2. Lack a consistent approach to performance management and career
development
- Don't help employees understand how these work at Google and doesn't coach them on their options to develop and stretch.
- Not proactive, waits for the employee to come to them.
3. Spend too little time managing and communicating
It doesn't at all follow that technical excellence isn't extremely important for bosses who haven't been chosen for technical excellence. (Perhaps it isn't; but Bock's work doesn't seem to provide any evidence.)