Why former employees say Microsoft can't innovate
betanews.com
betanews.com
Just to give the other side as somebody who worked there for a while and retired pretty successful, driving big changes isn't trivial but it isn't hard either. As an individual with an idea, you don't have to do ROI, fill out complicated budgeting forms, etc. Just make it work and then convince your managers -- all of whom are techies! -- that it's a good idea and you're willing to spend the time to make it happen. But I did see a lot of people fail with their "great idea" when they didn't want to be the one to build and ship it. Nobody there is going to do your work for you.
(Due Disclosure: I don't use any MS products and haven't in years)
You could argue that OneNote wasn't truly a bottom-up creation because Chris Pratley was already a pretty senior guy at MS when he started working on it.
To be honest, I'd have a harder time naming innovative features that shipped at Microsoft and didn't come up from the bottom. Pushing innovation from the top seems like it would be really hard for anyone not named Steve Jobs :-)
No one ever said MSFT didn't have top-notch developers. And developers tend to produce cool tools for themselves to use.
What about something more consumer facing? Has there been anything like that to come out in a while?
Another case would be digital music. After using partners to create MP3 players (which is their traditional strategy), they switched over to making their own (Zune). They just kept reacting to the iPod and iTunes Store. This meant they were playing by Apple's rules in the market, not creating the market.
What the iPod has is the iTunes store and household name (along the lines of Q-Tips and Kleenex). What Microsoft fails at is delivering an experience that makes an individual's entire experience easier. They are pros at providing tools, systems and support for the management of large deployements (e.g., corporate desktops/laptops, Apple and even Linux don't come anywhere close to the capabilities Microsoft provides in this regard), but they have always had problems directly targetting the consumer. The only direct-to-consumer product they have that they did good on in their own right is the XBox 360. They did this by being first and doing a great job of getting production companies on board for games. They did not go with the level of innovation that some companies did (PS3/Wii). It doesn't take innovation to lead the pack, or even just to succeed in an arena. It takes knowing who your target customers are and providing the experience that they didn't even know they wanted before they used it.
Even Apple in this regard wasn't very innovative. They took a concept that had already been proven in the marketplace (MP3 player), and made it better. I had an old Rio MP3 player back in the day, and it was a pain to get music onto it. When the first iPod came out, it was much more useful because it was easy to get my music on there. The innovation was in making it simple to buy music from iTunes and put it on your iPod.
I'm curious to see what the new Zunes look like when Microsoft announces them next week. But I suspect that it will be evolutionary, not revolutionary.
Hey, that's exactly what Google is doing in social networking.
The only teams at MS that seem to be succeeding are the ones not resting on their laurels, doing things the competition isn't doing.
Meanwhile look at the IE team - years and years of being bested by Firefox, and now being bested by WebKit, and they still don't seem to have their act together. Every single feature in IE8 was a me-too off of something Safari, Firefox, or Chrome already did.
Here's a quick and light article on the subject: http://www.joelonsoftware.com/articles/fog0000000070.html
Here's a video lecture on the subject: http://www.youtube.com/watch?v=FdkQwQQWX9Q
And here's the classical Harvard business review article that originally proposed it: http://cb.hbsp.harvard.edu/cb/web/product_detail.seam;jsessi...
Given the choice of doing exactly the same work for company A that believes that and company B that pays bonuses anyway, what sane person wouldn't go for B? That's not puzzling at all.
- Incentives are very hard to get right. If your bonus is tied to sales you inflate sales and don't care about margins. If your bonus is tied to margins you oversell and someone else suffers in aftersales support. If your bonus is tied to stockprice you do what you can to inflate the stockprice, hoping everything won't come crashing down before you receive your bonus. And so on.
- When it comes to it very few people are driven by money. They'll say they are but in reality they aren't. Would you knowingly sell something expensive to an old lady who can't afford it that she doesn't need in order to make a sell and get your bonus? Most people wouldn't.
- Incentives drive you to only care about the incentive for which you're awarded. If you're in one department and can make $10 knowing that another department will lose $100 that's what you'll do.
Besides very few choices in life, if any, are as black and white as you put them.
No economist believes that they are. People are motivated by a concept called "utility" which is, money, a nice office, interesting colleagues, meaningful work, a short commute, opportunity to assert authority, and other factors. The exact mix will vary from person to person, but the money is important: it often determines quality of life outside of work. All other things being equal, people will join the company that will maximize this, or appears to at hiring decision time. If one company has, even on paper, profit sharing/bonuses/stock options/whatever, and its rival doesn't, who's going to get the pick of the talent?
One simple and concrete thing that struck me was how the distance from the top to the trenches has severely increased as of late. The old metric of "no more than 5 levels" is being wildly violated by a factor of a bit more than 2.
Also high level overview on who was RIFFed in last year's surprise layoffs (hint, not much in the way of middle management).
Bottom line: What PG pointed out a while ago continues to the true and there's no signs that will change.
My answer to your point would be that it's the job of the CEO and board to structure the company so that it's as close to the ideal as possible. In the case of Microsoft, one approach would be to force less coordination on some of these separate units. As it is, they pay a terrific "architecture tax" by having to use and support (and generally only use and support) Microsoft products.
If it turns out they must do that to keep the core of Windows, Office and server products viable, well, that suggests they shouldn't get into the businesses that can't be usefully built on top of the core in the first place, and get out once they realize that.
But regarding one of your other statements, I am not so sure:
Harsh experience shows that an organization gets grossly disfunctional as you increase the number of layers from 5.
What about other organizations like Wal-Mart or the Navy? Surely there are more than five layers from door greeter or new recruit to CEO or 4-star admiral, yet the organizations are able to function and carry out their missions. Or are software companies different?
In software development, being a good programmer doesn't make you a good project manager, which in turn doesn't make you a good middle manager, which in turn doesn't make you good at making the business decisions of the management group.
Companies have the luxury of being able to do this in time frames not measured in minutes and seconds.
The US officer corps post-WWII also (in theory) defeats the Peter Principle by an "up or out" rule. If you do not advance in rank and responsibilities within set time frames, you are retired (which of course has its own problems).
I suppose the Microsoft "long at level" concept is related to this ... but I don't think it works in this field for some of the very reasons you noted.
I don't think most companies are different, at least when it comes to the non-rote stuff. I.e. in a fast food restaurant the people are the bottom are supposed to be following a "book".
With the exception of navies, I don't think that works in warfare (I make this exception since most of the lower people have clear missions like "keep the ship afloat" and "fire this gun"; that doesn't work for infantry, armor or tactical air (might work in part for artillery)).
It most certainly applies for software firms, e.g. the IBM Chief Programmer Team concept has been shown not to work in practice.
In both cases, the mission was hampered by political leadership. The military was winning the Vietnam War, it was just taking longer than necessary because the White House was micromanaging the war and establishing unrealistic rules and requirements.
As for piracy, the Navy has been kept out of the game because it's cheaper to pay ransom than to pay for any significant military operations against pirates. There was a time where we would say "millions for defense but not a penny for tribute", land the Marines in North Africa, and force the pirates to allow safe passage. Nowadays we'd rather pay the tribute--and invading Somalia would be a far more difficult mission.
One big leadership one, though, is apropos: the 6 month tour of duty "ticket punch" for officers. For the Army at large (at least for a long time) it was perceived that Vietnam was a side show to the greater Cold War. However it was also perceived that getting "combat experience" was going to be a necessary part of one's "resume" for higher advancement. So officers would come in for a 6 six month tour ... and I probably don't have to detail how badly that worked out.
What struck me the most WRT Vietnam was the learning part horribly broke down, at least in the Army (I actually haven't studied much about the Marines there).
One reason blue water navies tend to do better right off the bat in a war is that just doing the blue water navy thing (keeping your ships from colliding and sinking when they're out at sea per the doctrine) requires a minimal level of competence that just isn't required in ground force exercises (assuming you can get the money to do any of any scale).
Hmmm, gunnery exercises probably also work better. On the other hand, every major naval force at the beginning of WWII had major flaws with their torpedoes ... and the US Navy was pretty much the worst in dealing with them (Washington refused to listen and the various flaws had to be debugged in Hawaii; by contrast the Germans found and fixed theirs quickly and cashiered two officers responsible for them (the latter from memory)).
A Marine rifleman at the rank of private in a deployed MEF reports to a platoon leader who reports to a company commander w.r.t. a regiment commander w.r.t. a division commander w.r.t. the MEF commander. Private to general in 5 steps--and that general commands a full task force complete with air and support units.
The MEF commander reports to a combatant commander, a 4-star with authority over all military operations within a given theater. There are ten of these unified commands. (Sometimes a special command is created underneath the unified command, as in Iraq--the US Forces, Iraq command reports to US Central Command.) The combatant commander reports directly to the secretary of defense, who reports to the president. So the chain of command is around 9-10 deep for the entire US military, but about half as long for a task force.
Almost as importantly, just about every servicemember knows--and can tell you--his entire chain of command between him and the President. There's no ambiguity at all. You're fully responsible for everyone beneath you, and you are fully responsible to a single person above you.
It's probably impractical to organize a software company similar to the military, but militaries are very consciously attentive to organizational structure, and their organizational structure probably meets their needs more than most businesses.
Have you missed out the squad and battalion levels, or is the USMC organised differently from how it used to be?
And as for reporting to many people, I didn't experience that either. I had one manager, and was 6 away from Ballmer. My manager, skip level, and skip skip level were all in the same hallway as me and very accessible. My manager truly worked his butt off to ensure his employees were happy and headed where they wanted to head in their MS career (at least, as much as his power let him).
"Beg your pardon?"
"Eight bosses."
"Eight?"
"Eight, Bob. So that means that, when I make a mistake, I have eight different people coming by to tell me about it."
Doesn't the best technology come from well funded small companies?
Their corporate structure seems to be geared towards managing their complex, large projects like Windows and Office. With a vast project like Windows, it makes sense that they need teams devoted to certain specifics with middle managers. That structure keeps things focused on the goals coming from up top.
Unfortunately since everyone is focused on their little piece of territory, innovative stuff is happening on the outside and they don't realize until its too late. Then when they realize something is a big idea they toss it into the same machine they have rolling for huge projects like Windows or Office.
As HN readers know, resourceful startups can get a lot accomplished with a small team. If startups had a huge organizational structure with everyone at the top trying to get a piece of the action then nothing would get done.
It would do Microsoft a lot of good to have an internal YC-like program that gives resources to small teams with a good product idea, and then leave them alone until they had something to demo.
Edit: I read my last point and realized something - it sounds a lot like Google's 20% thing. Does Microsoft have anything like that already?
I'm not saying the article's points are not valid and true; as far as I know, they are. Just pointing out that if we're looking at what EX-employees are saying, we're looking at reality through a certain filter.
In a company where advancing means going into management, you do end up with what you are describing: no room for advancement, or too much management.
Thankfully there are companies out there where sticking to the technical track is a viable way to advance.