Gerstner saved a brand name by destroying a tech company and making it a consulting company. I'm not sure the same is possible with Intel.

I think that undersells IBMs continued technical achievement in the mainframe space. It's a niche business but the folks who need it really need it. Sure fine it's not their growth sector, software (thanks mostly to RedHat), but they managed to keep a really cool and unique segment of engineering going despite commodity hardware eating the world.

Gerstner left ibm far better than when he found it. That's the main point.

Now generally speaking most American OEMs lost their manufacturing chops since 1985 when software began eating everything.

Today, tech companies need to decide: did that go too far? Should we reincorporate physical r&d, and down size manufacturing supply chain bringing it back inside?

Those are legit questions. Summarily stating ibm is just a consulting company doesn't get us anywhere.

Positive counterexample? So you want Intel to go from a semiconductor manufacturing and design firm into a consulting and outsourcing company?

No. I want ibm to be a better ibm for its customers, and ditto for intel. Intel clearly is more about manufacturing than ibm. Put another way, dropping hw and hw manufacturing at Intel for something else is net worse.

IBM had three problems at that point.

The first was a culture of speaking to the expectations of senior executives, and following strategies that made no sense.

The second was that they were on all sides of all technology disputes - and so everyone hated them.

The third was that they were under an antitrust consent decree.

Basic software expertise wasn't going to help with any of this. The third was going to solve itself in time. The second was good old accounting. The first, well, https://gunkies.org/wiki/Gordon_Letwin_OS/2_usenet_post does a good job of describing what IBM was doing wrong.

By the time Open Source became a thing, IBM clearly had figured it out. The golden rule is this, people are willing to pay a certain amount for a solution to their problems. People don't care how their budget is divided. So if the complement to your solution is open source, you can charge more for the proprietary bit. Better yet, the open source thing with a bunch of support has a better shot than a completely proprietary stack. So open source the complement to your real product.

Gerstner somehow "got it". And it showed. This is why IBM supported Linux, Apache, etc in 1998/1999. Letting Microsoft lose the standards battles.

But do I see you saying that standards are a software issue? Maybe, but they also mattered to American Express. Which is where he came from. So he was equipped to understand the issue.

That Usenet post was a fascinating read. I’ve only encountered OS/2 once in the wild, and it was in a lab being used to operate a very specialized piece of equipment, and it was circa 2005 so it was all ancient relics. Oddly, the lab members spoke highly of the user experience in OS/2 but I never got to mess with it.

I learned a lot from that post. It is where I got a lot of ideas that I still use to understand the world today.

I have seen it at s friend when it was new. Multiple DOS boxes where nifty. I think it also was less prone to crashes from faulty applications.

You make good points, and provide good details. Thank you. I'll add around that. Based on a book or two I read on IBM at the time under Gerstner,

- talked to the top 50 or 100 customers to learn what IBM was good at, and sucked at. From this he learned that IBM's weird nomenclature and incompatibility with other LANs, disks, etc.. sucked. He learned IBM was a bit lost in its own world. Perhaps more critically written as entitled. IBM worked on fixing it. On the good side he learned customers like IBM for consulting/out-sourcing of projects. Whether apps, LANs, O/S, storage one at a time or all at the same time, IBM did it all and decently well. Customers had one guy to thank (when things good) or yell at (when things went wrong). Customer's prefer that over dealing with 4 different specialist companies they then had to corall onto the same page with attendant billing isssues. Insane costs for mainframes was another "IBM sucks".

- These points taken together is why Gerstner decided to not split IBM up, which is what the out-going CEO (Akron?) suggested. At the time specialist firms in printing, disk, LANs, etc were progressing faster than IBM. It was thought splitting IBM and selling off divisions might make those divisions more competitive and allow what remained to be more profitable.

- Learned that IBM had some good ideas that died in corporate American BS of meetings. Example: IBM had a manufacturing plan to reduce the costs of mainframes. When he found out never done, he got that going, and did something about removing the cruft of Corp America politics and nonsense that chokes out good ideas.

- IBM also got lucky that the internet era (think B2B) needed high capacity HW. Mainframes were key. They had the horsepower. IBM moved MSRP down, and made a lot of money at the same time.

- Gerstner was firm but fair. When he came in he said IBM's financial situation was dire. He said there'd be layoff's but once done, IBM would tend to not do more. So whoever was left was wanted, and was needed to make IBM better. That was both a like of remaining employees, a warm thanks to those who were left, and a warning, and an attempt to remove fear of firings in the future. Be here and be real or get lost. He sold off art, and real-estate to raise cash.

Also good points.

The layoff was straight out of Machiavelli. If you must do bad things, then do them all at once. If you can do good things, then do them bit by bit.

People will then forget the bad and remember the good.