[microarchitect]

Maybe I'm in the minority here, but I don't think "fixing" Yahoo ought to be as difficult as people make it seem.

I worked for a semi-dysfunctional semiconductor company and so I think I have some perspective on what it takes to "fix" a company like Yahoo.

I think the biggest reason a company becomes like Yahoo has less to do with senior management and more to do with the fact that the company is filled with low to mediocre quality middle managers and employees. Mediocre managers aren't very good at their job, so it seems like they make up by playing politics. Once politics is rampant, what gets done is not what works the best, but what helps you or your team get ahead in the political game. At this point you start bleeding talent because good people prefer not to work in an irrational workplace. You also lose the focus and initiative you need to keep a successful product successful.

The worst part is that there's nothing you can really do to fix this. You can try to fire the bottom 10%, but if you only have a problem with the bottom 10%, you're not yahoo. You probably need to fire the bottom 50%, but then you wouldn't have enough personnel to keep all your assets going AND some of that 50% would also be good employees who lost out in the last edition the political olympic games.

Any well-intentioned change in process or product focus by top-management will also inevitable morph into something completely different thanks to your low-quality middle managers who won't understand what you're trying to do, but will understand that they need to step up the politics to keep their jobs.

I guess the point of my rant is that a big company is like an emergent system and you can't really fix a dysfunctional emergent system by changing a small number of things that are part of that system.

[xyahoo]

You hit the nail on the head. Mediocre, petty middle managers are the algae that are sucking the oxygen out of Yahoo.

I may get jeered for saying this, but a non-trivial amount of it was a result of Jerry. He kept his buddies around in high places for too long, and the rot set in. There were VPs in Yahoo who contributed absolutely nothing (and probably are still around, but I left). But they were always blocking progress just to stroke their own egos. Once you have nepotism and incompetence at the top, it just flows down naturally and the next thing you know there is widespread politics and infighting between teams. It becomes a feedback loop; the good people leave, and are replaced with buddies of the current incompetents.

The way to fix this is to actually make the problem worse for a while, by segregating the groups into independent units. Then you fix each of these units, one at a time. Then you bring them back into the fold.

Can anyone there do it? I doubt it. Most of the decision makers are too busy playing politics.

[linuxhansl]

Yep that sounds about right. The other problem is that these "low to mediocre quality middle managers and employees" do not realize that they are "low to mediocre quality".

[schraeds]

"Swat Teams" of vetted talent to handle the most prized assets and initiatives. Let the cruft keep up day to day, apply proven talent to the most important initiatives piece by piece. ie. Apple and the iMac, then iPod, iBook, iWorks, Final Cut etc etc.

[demallien]

I actually work in exactly this type of swat team. We were all hired at well above standard payrates at my large company to build the next generation product. The response from the existing teams was immediate - they saw us as a threat, and started doing everything they could to make our project fail. In the end they succeeded in gimping the product by simply giving us interfaces that just couldn't do what a good product needed. The swat team was dragged into mediocrity. Now many of those rather expensive hires are actively looking to move elsewhere.

Moral of the story - building a star team in an existing company is a good way of building resentment about that team and is unlikely to breed success.

[PakG1]

This is why such swat teams need to be hidden or actively shielded. Motorola's Droid team was completely separate and able to stand on its own feet. The Google Wave team was based in Australia and had its own incentives and structure. Obviously, swat teams are hit and miss like any startup; they're essentially startups within a corporation.

Or you need to be actively shielded. I was part of such a swat team for internal tools in my first job out of university. The IT department did everything possible to kill us, get us fired, etc. We needed support from the CEO down to stay alive. In the days before I came on board, our managers and directors would disguise the team, telling other people, "Ah, they're just business analysts working on miscellaneous stuff. Not important. Ignore them."

[ichattopadhyaya]

I know why Flickr is doing so well. It is shielded and isolated from the rest of its parent organization.

[xxpor]

Why didn't they give you the power to demand things from other teams, lest they be fired?

[chime]

You can demand a functional API but you cannot demand a reliable service. That is what they've been unable to provide in the first place.

[demallien]

Well, there are several reasons. One is that this is France, and firing people is just more hassle than it is worth. Th second, and more generally applicable reason, is that that would involve changing the status quo. If changing the status quo was that easy, you wouldn't need a SWAT team...

[jleader]

When I left Yahoo in 2008, they had such a "swat team", at least for software development (they called it a "tiger team"). Unfortunately, I don't think Yahoo's main problems at that time were technological. Maybe they needed a similar approach to product management?