Hacker News new | comments | ask | show | jobs | submit login
Why can’t Microsoft get their products right on the first try? (owened.co.nz)
33 points by owenwil on May 30, 2013 | hide | past | web | favorite | 50 comments

Also ex-employee of Microsoft -- used to work in SQL Server (very traditional, old-Microsoft group).

Jobs people had in my group: (1) Developer: Write code. (2) Test: Write automation to test the code developers write. (3) Program manager: Make sure everyone is coordinated and that there's a reliable, up-to-date spec for the product.

Microsoft doesn't have a strong tradition of product management or design, and I'd argue those are the two areas where they're weakest as a company. Is there someone who works at Microsoft with the title "Product Manager"? I know there were "product planners", and "program managers", but not PM in the Silicon Valley sense of the term.

They're especially good at partner management though, and I don't think they get enough credit for that.

#2 wasn't always that. When I started at MSFT, many moons ago, there were also testers who solely provided end-user style testing and feedback on the product. While their loss was less traumatic to my own division (DevDiv), during and after that transition it seemed that replacing those STEs with people who wrote automation (SDETs) caused some of my peer product units to lose out on the only people with real customer empathy.

I guess in one way there are still STEs, but now they're called "customers."

I am a ex-employee of MS. In my observation, the decision making process is very complicated inside microsoft. Because of big-hierarchy, there are many people involved in making decision even for for a small feature. Most of the decision maker are non-expert on the topic. Sometime they argue: we should do A then after sometime few other manager will argue for B and again after sometime they will argue for A and this keeps on going till there is little time left for release. And at end, they randomly pick either A or B.

You're right, and it's a very disempowering culture for the 59s and 60s (these are internal levels) at the bottom.

I'm in SV now and no way I'll go back to that. Too much consensus, discussion, meetings, not enough autonomy, action, or customer orientation.

So, really, a few thoughts:

1. I think a theme that comes up on HN often is "when do you ship?". You could always argue that "well, we should wait to ship in order to [fix this/add this feature/...]". At a certain point, you have to say that you're done and ship it. There will always be room for improvement and the good news is that you're actually able to improve it later on. You just need to set a "ship" bar that is acceptable.

2. Like others have said, the first point is made even more important as Microsoft was already late to market with a tablet friendly OS.

3. People could argue all day about whether it was acceptable or not for the Office team to have released Office 2013 without Metro style apps. Whether it was acceptable for Windows 8 to be released before the Office team made Metro style apps. After taking points 1 & 2 into consideration, you have to remember that you need to manage resources. I'm sure the Office team WANTED Metro apps, but it was probably impossible for them to ship Office 2013 and Metro Office at the same time and "on time" for Office 2013 desktop release. Like I said, you could argue all day whether you think they managed their time/resources properly but either way there is something to be learned from this. You simply can't do everything at once. I don't know what the reasons are here, but for some reason Microsoft must have deemed it more important to ship desktop Office before the Metro apps. I'd also bargain that the Metro apps will have something to do with Office 365 subscriptions and IIRC the desktop Office 2013 release is largely testing Office 365 out (the consumer version, anyways). Anyways: point here is you can't do everything at once.

4. The article talks about perception as if it was a permanent thing. I'd say that perception can change without having it to be some colossal task. I have to go soon and the first example that comes to mind is people made so, so, so much fun of the iPad when it first came out. I remember people making fun of the first person I know to have bought one. And now? "Everyone" has one. I don't think Apple necesarily did anything to make this perception change, but after people saw the benefits of it they changed their attitudes towards it themselves. The same thing can happen here (e.g. if I didn't like Windows 8 but then I see someone using Windows 8.x in some way I think is really cool, it might cross my mind that maybe now Windows is in a better state and my perception of the product will change).

My $0.02

disclosure: i interned at microsoft in 2011 and 2012

Regarding 1: I think this is the product Microsoft wanted to ship. I don't think what they produced is necessarily a bad implementation of what they were trying do. I think their goal didn't match what their customers wanted. Some percentage(I'm guessing a high%) of users use Windows because its what they know and are comfortable with. Windows 8 throws out what a lot of these users spent a lot of time learning.

The bigger problem I think is that Microsoft failed to deliver not one, but two products. They wanted a new version of Windows, and they wanted an entry in the tablet OS market. Both are good goals, but not necessarily the same goal. In attempting to hit both, they failed to hit either one. Windows 8 is a bad replacement for Windows 7, it's missing features, and it's more difficult to use with a mouse and keyboard. Windows 8 is a lackluster tablet OS, it's missing a lot of the polish it needs to really compete with iOS and Android. I think Microsoft should have released not one OS, but two. Preserve application compatibility if possible, but don't sacrifice the ease of mouse and keyboard control at the alter of touch interface. A keyboard and mouse versus touch screen are different enough interfaces to warrant different UIs.

This isn't uniformly true, though. The original Xbox was, for all intents and purposes, and amazing product launch. Same with the Kinect, Office 2003, Windows XP, etc.

This models seems to hold more truth in recent years, however. I can't come up with a solid reason why this is, but it isn't an absolute truth.

Unlike the rest of their products, the Xbox wasn't a massive design by committee project. It was created by 4 guys from the DirectX team, effectively their own little startup within the Microsoft world.

The Xbox was also able to leverage the past ~5 years of DirectX development and PC gaming knowledge and built a OS that was very similar to Win32 minus the HAL.

I actually agree with you on this - but the Xbox came from a department that is segmented off from the rest of the company and uninfluenced by the rest of the corporate culture. The company made a breakthrough there, but in recent years they haven't been able to repeat that.

Am I missing something here, why is this "news"?

This "can't get right on first try" happens to virtually EVERY company. We don't even have to look too far back to find examples. Just one example would be Apple releasing Maps, which was oh-so-perfect out the gate.

I think the article is trying to say that for Microsoft, this behavior is the rule, not the exception. I was a bit surprised there was no mention of things like the Surface, Zune, Windows ME, etc. which all seem to fit the rule.

For Apple (in the last ~8 years at least) it seems to be the other way around, in that this behavior is the exception, not the rule.

Or even further back: Windows 3, Windows 98, Windows NT (early versions). There's a really really long history of doing that sort of thing at Microsoft.

Can't really compare to Apple releasing Maps though. They released maps because they had to, it isn't a feature neither a selling point. The contract with Google expired and the original Maps had to be phased out, also the YouTube app, and they had to come up with something. The app itself was superior, had good usability and cool features, but mapping data is largely a monopoly and you can't really compete without owning Navitec/Atlas/whatever.

They had another 6 months to update their Maps app... (released it in September or something, the contract ended in December).

They just wanted to take Google by suprise, so they wouldn't release their apps map on time and by this, they had upset their customers with a piece of crap product (Apple Maps :s)

Right, but even your example shows a reason why this approach is an issue. Apple Maps was late to the market - it came out after someone else (Google) already captured essentially the whole thing. Apple released a product that was maybe 25% of Google Map's feature parity and they were late to the party. They failed and now are trying to iterate out of a hole but their perception has already been set to "the inferior one" by most potential users.

But are there software companies that consistently deliver projects of the scale and complexity of Windows without hiccups?

Isn't the point exactly that Google and Apple are doing it with few major screwups? And in the near future we'll see Canonical and Mozilla's versions of mobile OSes, and see hoe successful they are.

Canonical is an interesting case. While all the others (except maybe Mozilla) do a lot of work on the systems programing, most of Canonical's system level stuff is done by many smaller, upstream, projects and Canonical is responsible for integrating them, and selecting ready for release versions. This allows them to much easier avoid to early to release problems, and direct their effort in a much more focused way on the end user experience.

Also, their duel release model (long term support and 6 month short term support) gives them a nice test bed where they can try out radical re-designs before forcing it onto the users, who can continue using the LTS release.

I don't think Apple or Google is doing anything nearly as complex as Windows 8, once you factor in the number of hardware partners and the constraints on backwards compatibility.

I still wouldn't call this "news" as I've heard this pointed out by numerous individuals before, but I think what he's commenting on is fundamentally different from Apple Maps. Maps' first release sucked, and they're obviously now working to improve it. On the other hand, other Apple products almost always "just get better", as do most other projects that come to mind. Sure, there are occasional bugs and gripes, growing pains, etc. but even when I used to be a big fan of Windows I noticed that they seemed to alternate between big wins and big fails. ME? Fail. XP? Win. Vista? Fail. 7? Win. 8? Fail. I'm less familiar with previous versions, but I suspect the general trend would continue fairly predictably. I'm less familiar with things like Office, so I can't speak from personal experience much there, but my point is just that Microsoft does seem to have a very strong alternation between fails and wins, which is different from simply a bad first-release.

It is very rare to get products right on the first try. There is a reason Google products launch as beta. Most companies go for limited launch of initial versions. Microsoft generally does a mass launch; that is probably the reason for the perception.

"No wireless. Less space than a nomad. Lame."

Sometimes chasing feature parity is a mistake.

And Microsoft don't really do that. They generally do best when they attack from a different angle.

von Clausewitz gave simple advice for winning battles: concentrate your strongest forces and attack at the enemy's weakest point.

How did Microsoft beat Borland? They attacked at the interface and ease-of-use, not on compiler speed.

How did they beat Apple on the desktop? They attacked on backwards compatibility, not on a better UI.

People that run big companies don't care AT ALL about getting products right. They want money, period. If they can get a promotion or bonus without getting the product right, they will. Sometimes that means stepping in front of progress.

"Announce Windows Phone 7. No multitasking despite all competitors having it, no application fast resume. No major applications that competitors have. No turn by turn directions. No front-facing camera."

Honestly I think Windows Phone 7 was a case of blatantly copying Apple. It came out at the end of the iPhone 3GS' life which had exactly "No multitasking despite all competitors having it, no application fast resume" and "No front-facing camera". Windows Phone 7 even threw away the ability to sideload apps which was a staple of Windows Mobile in favor of an app store only approach.

I think Microsoft made the mistake of thinking that they could pull off what Apple did with iOS, which is silly as Apple spent years building this unique position through the iPod. As a sort-of new entrant, Microsoft can't afford to start off with something that isn't ahead of the competition by a significant margin, no company in the mobile space other than Apple or Google can do this right now.

There also seems to be a lack in ubiquity across their products and platforms.

I remember when they had Windows Live Mesh, Windows Live SkyDrive and Windows Live Sync only to later merge them all into SkyDrive, then they had Silverlight for Windows, Mac and Moonlight for Linux along with Windows Phone 7's flavor of Silverlight and the XBox 360s flavor of Silverlight (used in ads) only later to can everything and now Windows 8, Windows Phone 8 and the XBox One all run the new Windows Runtime but don't have the same app store nor can they run the same apps for no apparent reason.

"Because right now, it's still behind the pack."

Um.. Active Directory run's pretty much every Enterprise and Corporation worldwide. From that, I'd say Microsoft "has got it right".

Active Directory is pervasive because Windows is pervasive, and Windows is pervasive because MS Office is pervasive. To my knowledge, there is no other effective central AAA (authentication/authorization/accounting) solution for Windows than Active Directory.

Cloud-based office productivity products are slowly chipping away at AD's necessity, but there's still a need out there for a cloud-based AAA solution that meets the needs of BigCos and SmallCos alike. Unfortunately it's going to be a while before we get there because it will have to emulate AD in the meantime.

> "Active Directory runs pretty much every Enterprise and Corporation worldwide"

From that, I'd say: so much for "we're a devices and services company now".

Microsoft ships minimum viable products then iterates based on feedback. In the software industry, lock-in effects can give a big advantage to first to market.

Well yes, and unfortunately for Microsoft they are never first to market. In my opinion, the iteration model only works if you're dominating the market. Otherwise, you need to create something that blows your competitors' products out of the water.

It's hard to know what will work before real consumers have had it in their hands for a while. Perceptions change with propinquity.

Microsoft is one of the few companies that literally define modern consumer computing. It makes me very happy every time I see them moving forward. Sometimes it works, sometimes it doesn't. When it doesn't, they grow and fix it.

Apple has the same cycle, but it's less immediately apparent.

For me modern computing is defined by Apple and a bit by Google - I don't knowingly use a single Microsoft product (who hosts iCloud stuff - is is MS?). How does MS define modern computing - are you meaning in the last 10 years or now?

Microsoft owns about 90% of the desktop market, 30%+ of the server market and 25%+ of active gaming consoles. Right now.

If you could exclude work computers, and include mobile ones (phones and tablets) that would look a bit different. Possible bias: My circle of friends and colleagues is obviously seriously skewed as MS is a small slice of the pie in my little world and that of those I know.

It could be that there is a very large disconnect between those promising features and the team that ultimately decides on whether or not something is good enough to ship.

Let's say that the front office (those doing the selling) start promising tons of features because they want to really sell this thing. Perhaps those features might be a little more complex than the "visionaries" first imagined. Now the expectations have been set.

The engineers try to meet those expectations. They try their best, but then it gets to QA. The people doing QA really care about their job, about the quality of what is being released, and they do a good job poking holes in things trying to get it to a quality that was promised. QA sends it back, but the developers have moved onto producing new features. Bug fixes are not cool. Sometimes doing it right is hard. Things start to slow down.

The end product is something that doesn't quite meet the expectations and is late...just a guess though.

who can really? (insert quote of linus torvalds awesomeness here) their problem isn't that they can't get it right. their problem is that it takes them forever to fix it.

Microsoft's "the third is charm" strategy is wldely known for year and tolerable since consumer is pretty much locked in by office software, and Microsoft knew about this.

It is until apple spoiled consumer with perfection and true use friendliness that made the contrast that Microsoft product along with the "beta" testing with paid product is no long bearable.

Still, Microsoft can afford this lossy strategy because they knoe nobody can really stop using office.

Personally I don't think Microsoft's primary agenda is to produce user-enabling products. Imo their desire is to produce user-hobbling products; products that slow down or preferably reverse the knowledge-empowering potential of computing machines for the average person. It's a political not marketing objective, and you'll never see them admitting to it. You have to analyze their actions not their words to see it.

If you assume Microsoft has always been trying to achieve the 'best' product possible, then their development history looks like incredible incompetence. If you view it in the political context of Elites trying to cripple development of 'computing power to the people', it makes perfect sense and reveals a high degree of sophistication in the Microsoft inner management group.

Effectively their strategy is to steer development of their products in the most socially harmful direction possible, constantly pushing the boundary of what the market will reject as 'too stupid'. When they hit public resistance, they wait as long as possible then back off with a slightly less stupid product release. Once that version has become entrenched they then try again with something even more stupid. Windows 8 was just an example of MS pushing a little too hard.

I am deeply bothered by this idea that you suggest and supposedly hold yourself. It essentially "presumes malice" instead of assuming ignorance. To believe that their desire is to produce user-hobbling products is to assume that somehow there is a conspiracy of executives at a company whose explicit policy is to do harm. That somehow no one has revealed this fact, their board has been somehow complicit in this direction as revealed to them by executives, and that they sit in these, perhaps smoke-filled, board-rooms discussing how they can dumb down their products or harm the general public.

That's insane, you're insane, and your philosophy is far and above more cynical than anyone's need be.

Temporarily ignoring all of Microsoft's specific problems (and there are many) and any other company's specific problems for that matter (every company has them)...

Nobody gets a product right on the first try. Full Stop.

The more interesting question here is this:

Why does Microsoft expose the first try to the public?

I can't help but think that Vista -> Windows 7 is an example of this pattern.

    Win3  > Win3.1
    Win95 > Win98
    WinME > WinXP
    Vista > Win7
    Win8  > ? (unlikely to be 8.1)

Was Win95 a failure? From what I remember, it was a huge, smashing success for them.

Wasn't 95 the first "full" DE and sligthly better game support? With 3.1 you'd still boot into DOS and start up windows to play solitaire.

DOS was still there under the hood. It wasn't until Windows XP that they officially killed off DOS (there was a cheesy skit about this at the launch event with Bill Gates killing off DOS).

And suddenly old game compatibility was much iffier. Good thing DOSBox stepped up.

WindowsME was a band-aid. It wasn't meant to exist. Windows2K was supposed to be the release that merged the Win9x and WinNT worlds. It didn't end up working, so Windows2000 was released as the next version of the WindowsNT line (WinNT 5.0) and Windows XP (WinNT 5.1) did what it was supposed to do.

There are probably some parallels to Longhorn/Vista. Except it seems like with Vista Microsoft bit off a lot more than they could chew, and had to massively scale things back for release.

... ME -> XP -> Vista -> 7 -> 8 ...

Do they care to?

Seems like they make gobs of money regardless.

I was going to say that same thing...their strategy seems horrible to the technocrats, but they continue to push on areas until they are a major player. I'm no Microsoft fan, but you can't say their strategy isn't working for them...Windows, Office, and Xbox are all huge in their respective markets.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact