Hacker News new | more | comments | ask | show | jobs | submit login

IBM is suffering the fate of OPEC countries, except with technology. IBM wasn't diversified. (Sure, IBM has lots of products, but it usually sold licenses to all these products in large bundles, rarely did anyone buy just DB2, rather, agreements purchased entire ecosystem of IBM software/hardware). IBM was built on a single golden goose "massive multi-million dollar agreements with Fortune 500 CTO's for thousands of licenses of IBM hardware/software". IBM solved big problems with big, expensive solutions. Nowadays, developers regularly make the purchasing decisions, not upper-management, so FOSS or Freemium products win out over expensive IBM software licenses. IBM hardware is largely dying out because the Cloud and x86 won the architecture wars and no one wants to develop for PowerPC or AIX. So the legacy IBM is dying on the table.

IBM is losing its old accounts and old business agreements faster than it can generate new revenue with Softlayer/Bluemix/Watson. IBM makes 4.5B from Cloud revenue. In contrast, total revenue in 2015 was 81.7B. So IBM revenue is going to be a bloodbath until Cloud revenue can catch up. Unfortunately, Cloud is a razor thin margin business, where Amazon, Microsoft, and Google are willing to sell at, or below cost to gain marketshare, which will likely continue to hurt IBM's cloud revenue at a time it needs it most. IBM is the underdog here, so it can not bring any of the other cloud giants to the table and make them end the price wars, so IBM cloud revenue is likely to remain anemic.

IBM also suffers from a number of management mistakes. IBM is way behind on automation and lacks the talent, so they've had to make acquisitions and "strategic partnerships" to buy it. IBM previously tried to outsource as much software and operations to China or India, and the unskilled labor available in those countries simply don't have the skill sets to do devops or any sort of basic automation. So you've got datacenters where people still ssh into each box and run commands by hand thousands of times because the skill level isn't there. But hey, a few quarters of lower labor expenses in exchange for long term growth is a great deal when you can just parachute out.

Other mistakes IBM has made includes massive buildup of debt and stock buybacks to inflate the stock price. As bad as the stock looks now, just consider how bad it would be if IBM hadn't spent 8B+ buying back stock. These buybacks have largely been financed by new corporate debt. IBM's market cap is nearly 40% less than only a few years ago. These buybacks hide a lot of the stock price damage (the stock would be worth about $75 instead of ~$130). In addition, buybacks are essentially returning money from the corporation to the shareholders, unwinding the assets. This means IBM management has no plans which could return a higher yield to investors than simply handing back their money -- not great for investor confidence or long term future for IBM.

Source: Employee.

IBM's trick was to recognise that nerds are bad at selling stuff. Even up to the 90s, sales + support + okay product > great product. Then nerds took over IT (the world?):

> Nowadays, developers regularly make the purchasing decisions

Oops. Suddenly, you can't just cover over product weaknesses with sales. I want a slick, downloadable, preferably OSS demo which I can evaluate (and fix), not a slick sales guy. They just lie.

> IBM is way behind on automation and lacks the talent, so they've had to make acquisitions and "strategic partnerships" to buy it.

... and then they constantly scare away that acquihired talent by e.g. not using `git`, but their own SCM called "Rational Team Concert" (which is worse than `svn`). Suddenly, the avg. day of a dev became 1/3 as productive and x5 as frustrating.

It isn't that they treat developers horribly, they just don't know how to please them. You end up with dev teams of barely average "enterprise" developers, who are put under pressure by flailing managers and executives to "deliver". So they simply add to the decades of technical debt. Now most products are "rotten" to the core in this way, so you need ever bigger support and manual test/QA teams. Release cycles elongate, features take months to develop in a fragile code base. Sales, assuming they weren't douchebags to begin, have nothing good to sell. So management "incentivises" them instead of investing that money in a good product. Thus, the cycle continues.

In this way, it's also insidious for your career unless you want to move out of tech (because IBM's general reputation is okay). If you want to stay in tech you really have to be keeping up with the outside world, otherwise you'll get a nasty shock in that interview when it turns out you don't know how to use `git`.

    > faster than it can generate new revenue with Watson
I evaluated the Watson public APIs a little while (a year?) ago, and found them close to useless, yet the amount of hype behind the brand is intense. I wonder what the disconnect is.

The speech transcription API is pretty great, but that's the only one I've dealt with in any capacity.

IBM is huge with hardware, software, global services and other biz units that makeup the entire picture. They try and put on this face of "were so advanced" with watson and buying companies to stay ahead but in the end it's old school management that fails the company. Global services is the majority of employees and middle management is so bad they couldn't distinguish good internal talent if it hit them in the head. It's all about supplanting highly skilled labor with offshore talent, your job is truly replaceable. I've seen $200+/hr rates get outsourced to multiple $25/hr unskilled labor. Ask any major company who has brought IBM in and you'll hear horror stories. If you're an employee of a Fortune 5xx company where IBM comes in to take over your I/T department, RUN!! They care about money, not people.

IBM is easily replaceable with hardware, software, and finance solutions and you have young I/T managers who could give a rats a$$ about the old sentiment "buy ibm and you can't go wrong". Google, facebook, amazon, yahoo have all shown you can build and scale solutions beyond anything that a mainframe can do and much cheaper. IBM keeps living in the past and the most messed up places of employment I've encountered.

agree with everything @afforess says.

Unfortunately, this is 100% true...

Does AWS approach enterprise sales the same way IBM does? I know IBM offers the one point of contact approach (contact your rep; she/he takes care of it).

Yes, this is one thing they do fairly well.

> So the legacy IBM is dying on the table

If it is dying, it will be a slow death. IBM spent decades entrenching itself in government and Fortune 500 enterprises.

The 23 year old Ruby-on-Rails coder sitting at his desk in his unicorn's open floorplan office in SoMa may not be aware of it, but there are a ton of IBM machines out there in the world right now, running programs written in JCL, Rexx, Cobol, PL/I, and Fortran.

The first z196 mainframe to roll out six years ago went to Citigroup ( http://www.informationweek.com/it-leadership/global-cio-ibms... ). With the pressure for increasing quarterly profits and other priorities, why would Citigroup decide to spend a fortune to upgrade their ancient legacy systems, hoping that there won't be too many bumps along the way? IBM makes it easy for them to keep their legacy systems around - it's much cheaper, and it works. The systems have worked for decades actually.

When JPMorgan Chase decided to start cutting down on their IBM usage twelve years ago and start doing more IT in house, they had to strike a deal with IBM to hire about 4000 IBM IT workers ( http://www-03.ibm.com/press/us/en/pressrelease/7294.wss ). Since JPMC just put up a job posting ( https://jpmchase.taleo.net/careersection/jobdetail.ftl?job=1... ) looking for someone with "Expert Skills in IBM-Z/OS COBOL, CICS, VSAM, JCL, MQ, DB2 SQL, Assembler and Stored Procedures", I'd guess some of those systems are still in place.

Note the location of the job. Tampa, Florida. Not exactly the center of what's young and hot in the tech world. More like the kind of place an IBM old timer might go to retire, before he decides he was bored and would like to stick his hand back in and get some work. This JPMC job ( https://jpmchase.taleo.net/careersection/jobdetail.ftl?job=1... ) looking for "Experience with z/OS ISPF, JCL, and JES2" is in Bangalore. Not the center of tech, but plenty of people out there are working on these mainframes and with languages listed in the job listings like JCL, REXX, Cobol etc. These jobs are all at a bank which tried to start weaning itself off IBM twelve years ago.

Governments and Fortune 500 companies have plenty of IBM mainframes out there running, often with code written half a century ago in PL/I or the like, running business processes which are still used, and they're not going away any time soon.

The problem is that being "the legacy vendor" is not a good place to be. Your margins can only get lower. You get a huge runway towards building alternatives, but building alternatives you will have to do.

Applications are open for YC Summer 2019

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