*IBM: The Rise and Fall and Reinvention of a Global Icon*

That is the new and excellent and I am tempted to label definitive book by James W. Cortada.  The author worked at IBM for thirty-eight years, a reasonable qualification to attempt such a tome.  Here is one excerpt:

It is difficult to exaggerate the importance of the Social Security win to the evolution of IBM.  That one piece of business, along with its effects on other agencies and businesses, wiped out the Great Depression for IBM.  That transaction handed IBM a potential market of 20,000 other companies that would need to process social security data.  When the books were closed on IBM’s business in 1937, revenue had increased by 48 percent of 1935’s, and by the end of 1939, by 81 percent of 1935’s.

And then for the 1960s:

IBM’s System 360 was one of the most important products introduced by a U.S. corporation in the twentieth century, and it nearly broke IBM.  A short list of the most transformative products of the past century would include it…

On April 7, 1964, IBM introduced a combination of six components, dozens of items of peripheral equipment, such as tape drives, disk drives, printers, and control units, among others; and a promise to provide the software necessary to make everything work together — a mindboggling total of 150 products…manuals describing all the machines, components, software, and their installations and operation filled more than 50 linear feet of bookshelves.

But later on, by the 1970s:

With ten layers of management, each with staffs, it was probably inevitable that bureaucracy would grow.

The research and background context is amazing and the book is readable throughout.  You can pre-order here.


IBM got paid by both the Axis and the Allies, as each started to expand their government services. Watson was some kind of evil genius. Too bad, Microsoft, its former OS partner, and Apple, its barely-on-the-radar consumer competitor, both surpassed IBM and each became the world's most valuable company for a time. IBM unfortunately today has a bad rep in tech circles as a dinosaur lacking in vision. If you have any kind of talent, you work at FAANG or maybe high finance, not IBM. See kids this what happens when the government is your #1 customer. Engineering dies while compliance drones who know how to navigate obscure government procurement rules take charge.

In 2017 IBM was granted over 9,000 patents. The most of any company in the US.

But keep spewing your anti-govt screed.

And yet IBM's stock performance has been seriously lagging the market for years, which suggests that their patents aren't very valuable.

Patents aren't the best measurement for innovation and is these days increasingly a sign of rent-seeking, which describes IBM to a T. Here's a few recent ones:


Patents should have not have been granted for some of these. Too obvious, like 9,564,919 for "Managing Data Records":

"In response to determining the estimated record length of a compressed second data record is outside of the amount of free space in the page, the database management system may determine whether an estimated length of a compressed page is outside of the page length. In response to determining the estimated length of a compressed page is within the page length, the page may be compressed."

What a joke. This extremely obvious "ingenious" idea of compressing data so it fits the bounds checking constraints of your data store is used a million times over in FOSS software. Prior art should nullify this government granted monopoly to big business. It took 6 guys in Beijing and 1 guy in CA to "discover"this? IBM is clearly a bureacratic monster of make-work and fake-work. USPTO needs to hire guys that really understand technology so they don't get stinkers like these through the system or a big company will just bully small companies with legal threats based on fictitious innovation. Policy wise math and software patents should be never be granted.

As a programmer at the time, I looked at the features of the first IBM PC DOS and said "Why did they buy this? They have people who could write it in a weekend?"

They did. It was amazingly a voluntary choice to give away Microsoft's ultimate wealth, because the structure of a semi-skunkworks project demanded it.

More generally though, technology translations are hard, and the number of top companies that jump them are very few.

Man am I old. I learned Fortran on the IBM 360 w/a WATFOR compiler. Staying up all night debugging and eating Shakespeare’s pizza. Them was the days.

I learned Fortran on an IBM 1620 on a three day course add-on to an electrical engineering degree in the mid 1960s.
It made self teaching BASIC in the early 1980s that much easier. My home computer with a massive-for-the-time RAM of 32K with keyboard and television set VDU was light years ahead of the 1620 which used punched cards for input and electric typewriter for output.
As far as I know IBM is now a company the bulk of whose income comes from software.
I don't think the "machines" part of "International Business Machines is relevant, although their research seems to produce stories about artificial intelligence.

Are your skills worth anything or does tech chew up its own workforce? Only asking because DST Systems laid off a bunch of COBOL programmers last year and that got my attention since conventional wisdom was that COBOL will never go away, a job for life.

Man am I old(er). In 1965 I was programming IBM computers in machine language,

The Shakespeare's in Columbia MO?

In the early 80s I worked for a startup in Austin. At the time, IBM religion was SNA. IP/Ethernet was just rearing its head. Since IBM refused to make an Ethernet interface, there was a space for IBM to Ethernet interfaces. My little outfit filled it. (As an aside, there was a micro power struggle in our little world for who would be the SNA expert. I lost, and was thrown a bone which amounted to forcing me to learn IP, Ethernet, and Unix. Oh alas.)
Anyway, I have the memory of a meeting with IBM guys, white shirted and tied, us in jeans, about a govt contract where they really needed an Ethernet interface. Their lead looked across the table and said, "You'd better be able to code."

And IBM found token ring a much better LAN technology - which was basically true until the need for better networking forced Ethernet into becoming a reasonably useful technology.

The decade of the 1980s was the decade of IBM as companies shifted to PCs from mainframe computers and everyone was provided a computer at his or her desk. Today, a computer on every desk is so common that younger people must assume it was always that way. No, it wasn't. And to lead this revolution, IBM hired an army of sales personnel and support, IBM often the biggest recruiter on college campuses. There was even an IBM look, white shirt and dark suit, handsome young men and pretty young women, selling an enormous volume of PCs and peripherals to companies hell-bent on being at the leading edge of this revolution. Of course, the revolution came and went, as all revolutions do, as low cost competitors rendered hardware an unprofitable business. And the army of sales personnel at IBM soon receded to little more than a few holdovers to talk about the glory days of big ticket sales, commissions, and bonuses. The number of IBM employees collapsed in the early 1990s, and only recently has IBM approached the peak in the 1980s. On the other hand, as the number of employees collapsed, revenue per employee exploded, as IBM shifted away from hardware to software and consulting.

Actually, things have returned to the early days in key respects. Before workers had standalone PCs, they used terminals with the storage and processing being done by the mainframe. But the standalone PC model has been in decline for 20 years, and people increasing use devices where storage and processing are again done centrally by cloud servers. This includes smartphones (that voice recognition isn't done on the device itself, but in the cloud -- which is also where your photos go for permanent storage and where all the GPS map data lives). Schools are switching to Chromebooks in droves, where none of the applications or user data are stored permanently on the device (which is what makes them extremely easy to manage). Vendors like Adobe (Photoshop) and Microsoft (Office) are pushing a cloud-based subscription model for their software (Adobe no longer even sells a 'shrink-wrapped' version of Photoshop at all -- a cloud subscription is the only option).

The path of IBM is like the path of American business, shifting from large employment of makers of hardware (IBM made those desk tops) to relatively small employment of knowledge workers; the large staff of sales personnel were the most visible. Consider how many fewer employees it takes for the cloud servers. IBM has evolved into an international company with a large presence in Asia as well as in the US and Europe, yet it has fewer employees today than in 1989. The reinvention of IBM, from mainframe to PC to software, consulting and cloud service, has dramatically changed not only the number of employees but the skills of the employees. Several years back GE tried to reinvent itself as a tech company that is also an industrial company. industrial company, deploying tech to increase the efficiency of manufacturing, which I thought had great promise (both for the company and the American worker). Alas, it did not work, at least not thus far. Successful reinvention seems not to include the vast majority of American workers. That's not a complaint, simply an acknowledgment of reality.


I felt a bit aroused when I saw IBM is supporting a developer community for quantum algorithms. https://qiskit.org/

Of course, they want externals to find applications (for free) to their shinny new quantum computer. If you put in the shoes of the guy who can code some real solution in that machine.....you won't have a patent that makes you crazy rich, but a salary of 300K is waiting for you.

The hardware for quantum computing is still decades away with current approaches. Still fun to think about the algorithms though.

Thinking about the code for a machine that does not exist yet as Lovelace and Babbage. Quite romantic issue.

IBM has working quantum hardware right now. The public can even sign-up to run experiments.

One has to wonder if anything would be different had the Department of Justice not been hounding them for so many years for being a "monopoly". You could see many times where they held back against competing against Microsoft even though they had the products to do it. My hunch is that they were afraid of the DOJ accusing them of being anti-competitive.

Today, IBM is just awful. They buy other software companies, cut off access to product documentation, which then forces their customers to hire IBM consultants or their partners' consultants to install and maintain it. I have no idea how this model is working out for them, but wow it sucks to work with their products.

You're probably right, as to bundling fears, but it also could be IBM's arrogance in believing it could license from this small company, Microsoft, and later develop its own operating system. Stupid decision.

The stock market obviously does not think that IBM has "reinvented" itself yet (and seems dubious that it will) so this book sounds premature.

James Cortada has written some definitive books on the history of computing. I would recommend his "The Digital Flood" for those who want the details on the spread of computing.

Cortada is somewhat partisan when it comes to IBM (he was an IBM sales manager for many years). So expect a rose tinted view of what IBM do.

Of course, there are other historians of computing.

What, no comments about how IBM at least survived, when its once-nearest competitor, Digital Equipment Corp., became irrelevant and then disappeared?

Overall the attrition rate for computer hardware and software companies has been ferocious ever since the days when Eckert and Mauchly decided to build one.

Except for IBM, which, for all its faults, seems to have learned how to at least survive even as most of its competitors over the years are now just memories (or perhaps a footnote in some obscure history-of-technology book).

The reason that OS 360 about sank IBM is that the software cost more than the hardware.

I'll probably not read this book but if it doesn't mention the predominance of former salesmen in the higher strata of IBM's management it is worthless.

In my eleven years in the late 50s early 60s (all as a finance guy in "IBM World Trade," its ridiculously-named international HQ company) I encountered many former computer salesmen who had been placed in charge of departments (personnel, finance, etc) for which they were not qualified.

Example: the ex-salesman in charge of international finance once "forgot" that a loan installment was due and when he was informed of the oversight acted the way any salesman would: he yelled at his subordinates.

This "successful salesmen can do anything" nonsense was introduced by the Watsons (father and two sons, one an alcoholic) is largely responsible for its management failure. IBM once (early-mid 60s) "owned" the computer/data processing industry and with smarter people at the top would have maintained that lead, being X times larger than today's company. But the "100 percent club" alumni gang in charge was too dumb.

Data hygiene services provide good data management strategy is also necessary if your company plans to analyze this data in order to extract actionable information.

Comments for this post are closed