Under Ballmer, Microsoft had the same pattern repeatedly: Zune, Kin, Windows Phone, Surface, Windows 8 -- they took too long to see the need, took way too long to make the product, and when they did it was either not good enough, too late, or both.

This is completely irrelevant. By the same logic, iPod was nothing special, just another mp3 player.

IE and XBox managed to muscle into an established market. This is harder than making a new market, and should very much be congratulated.

> In fact with IE 6 he so badly misjudged the market that he assumed the game was over and didn't even bother continuing development. IE 7 didn't come out until 4 years later!

Not so much misjudged, but started seeing as a threat. IE development hit a wall because the meme that web-based apps will replace windows-based apps got popular, and Microsoft didn't want to wipe out their own monopoly. IE4-IE6 were, at the time, absolutely crushingly good. IE6 is now remembered as the really bad browser, but that is because they stopped updating it, and in the later years it had to compete with much, much better browsers.

> You count XBox as a success, yet MS is in the hole over XBox by over a Billion dollars and will probably never make money on it, taking past losses and write-offs into account.

XBox also supports windows game development, strengthening their main money maker. What is that worth?

Microsoft did a complete re-write to componentize the browser and after IE3 (I think), IE was generally better than Netscape. At least, Walt Mossberg said so.

Hard to know why Microsoft stopped developing IE6 and you may be right. However, the next browser was scheduled to appear with the next OS, which was Vista, which turned into a debacle. If Vista had come out on time then it might not have been so bad.

You will recall that the US gov sued Microsoft on the basis that the browser was a separate market from the OS, whereas Microsoft argued it was a part of the OS. This gave Microsoft no incentive to develop IE on the sort of separate fast-track schedule that would have been best for the industry, and for consumers.

That's probably just one of the ways the anti-trust action helped harm consumers. Another was that it gave OEMs complete freedom to install crapware, which also harmed the whole PC industry.

Surprisingly little known story: in short MS was afraid of Linux as desktop OS, OSS world was seen as capably of quickly copying any innovative single feature. This is resulted in idea of "integrated innovation": produce innovation in app development, app framework, data access, communication, os based on it and ship it all at once, so Linux would never catch up (.Net Framework, Avalon - WPF, WinFS, Indigo - WCF, Longhorn). Obviously, plan failed (well, Linux failed to come to desktop too).

So IE was seen as done and future to be elsewhere. After IE6 ship with XP, IE team was transformed into Avalon team, consumed bunch of other teams and at it's peak Avalon project had something like 500 people working on it (after much struggle and numerous delays it shipped as WPF). So for something like two years, starting at summer of 2001, only 4-6 people worked on IE - IE Sustained Engineering team, patching security vulnerabilities. Then it became obvious that they can't keep up and some rearchitecture and refactoring are required in order to fix security problems for real. So IE Security Team was created, and some old timers were transferred from Avalon to do this security fix up, probably around 20 people.

Then it became obvious that "integrated innovation" is, in fact, too integrated to ship on time, Avalon and Longhorn are nowhere is sight, and browser war is effectively being lost for the second time to Firefox, of all things (MS management couldn't take fork of OSS fork of NN seriously for awhile). So IE Security Team was expanded, became IE7 team and happily started to work on a browser again, 3-4 years too late.

>However, the next browser was scheduled to appear with the next OS, which was Vista, which turned into a debacle.

As you can see from the story above, initially it's wasn't scheduled to appear at all.

Presumably all that happened under Allchin after bradsi left....

There is another twist - after XP Windows was split into two branches (since they didn't want to delay client to wait for server): Server and Longhorn. Brian Valentine managed Server (shipped as Server 2003), at same time different war room, different managers managed Longhorn ship. So essentially experienced management team was shipping server, and non-so heavy weight - Longhorn. Then some people were taken to manage XP SP2 (result of "security push"). So there is no surprise Longhorn had problems and required "reset" to finally ship as Vista.

And the lesson is: never ever branch to v+1 and v+2 branches unless absolutely required - one of them is going to be neglected.

IE6 was a terrible browser, but it was the best browser at the time. The anti-competitive behaviour helped, and then of course they sat on their laurels for years, but the best browser won.

Everyone seems to forget the most important examples...

If you forget that they now have 12 or so divisions with $billion plus in revenue.

But no product lines last forever - even Windows is bleeding market share, and upgrade rates are dropping like a brick. It seems a lot of people are happy with Win7 (or even WinXP) and not moving.

This strikes me as a Blackberry situation - market leader by a mile, but attempts to modernize and develop the Next Big Thing have all failed. They have enormous momentum and can coast for a long time - but what happens after? RIM failed to react to changing markets due to iPhone/Android, and arrogantly threw around their (momentum-based) sales numbers. Then those dried up.