Column A fun evening's entertainment pre-COVID was to find a pub near a large corporation's IT HQ, look for the customers with the haunted, desperate eyes, and ask them gently how "the migration" was going. Didn't matter which company or what migration. They're hard. They go wrong.
Not all migration misery is equal. Imagine you are a global IT brand, trading on the days when you were the undisputed heavyweight champion of the world, selling other enterprises services and systems based on your unquestioned competence. Imagine further your internal email system has broken so badly that your employees can keep quiet no longer, and have gone public in frustration and despair. Imagine you are IBM.
IBM's internal email failure is of Shakespearian richness. Tragedy, comedy, pomposity pricked, the decline of a great house, yet greater disaster foreshadowed. We don't know the details, because the company has merely acknowledged problems without commenting on nature, scale, or prognosis. We don't know, but the spilled beans form a pattern.
In 2017, IBM offloaded the email systems it owned – Notes and Verse, which sounds like a column in a 1950s literary magazine – to Indian company HCL.
IBM used those products itself, as you're supposed to if you're a global IT company selling email. But after a bit it got uncomfortable with all its corporate emails living away from home. So IBM decided to build its own infrastructure, take back control. And at the moment of migration, the 18-month-old project failed. Oops.
What went wrong
So many questions. Let's start with why IBM, the quintessential enterprise services company, sold off its enterprise email offering. No doubt it was thought that the billion-and-change deal would offset the RedHat $30bn+ deal of the same era – but that's not it.
IBM wants growth, and corporate email – especially a stagnant dinosaur like Notes – grows about as fast as Brian Eno's hair. Flog it off, cut the expense of supporting it, add the sale price to the bottom line, and investors love you. When the thing you're getting rid of is a core part of your company, though, the fun can get intense.
Why did IBM take so long to create local infrastructure for its email? That's harder to explain, except through the magic of mismanagement.
It could be that a simple plan grew out of scope, with the "opportunity" to add in the company's Outlook services for a single managed email backend.
The complexity of the task may have escaped the architects, or an initial plan may have proven impossible when a key component just wouldn't work in the new environment. Either option means many restarts and project prolongation. It could also have been underfunding, corporate politics or just plain incompetence. All of the above have sunk migrations.
There are two further factors that might have sunk this one. The first? Migrations that happen because of business units being sold have the hardest of hard deadlines for moving stuff that now belongs to one party off the infrastructure belonging to another. Security and compliance is nightmarish.
The second factor is that IBM is IBM. If a bank decides to migrate a massive on-prem farm into the cloud, it can choose between vendors. If IBM chose AWS or Google or Azure, what would that say about its own cloud? Not an option. So you don't always get to pick the best tools for the job – often you don't get to pick at all. If IBM Cloud were best in breed for this, then that wouldn't be a problem. Hands up if you think that's the case.
Could you take your eye off Wall St for a sec?
All these potential issues we've speculated about are just symptoms. All could be foreseen and planned for, made part of the decision process. IBM used to be good at foresight, planning and decisions. The reason IBM became a behemoth of IT in the last century was due to a combination of factors that relied on this: ruthlessness in understanding and manipulating its clients, a corporate culture that bordered on the militaristic, and a fierce focus on engineering for purpose. It was unbeatable for a while. It took revolutions in commodity silicon and software tools to overturn.
Uncertain in a world where it no longer set the rules, IBM swapped a focus on technological engineering to financial finessing. It was – is – a big and rich company, so by flogging off assets and outsourcing expertise, it could always find a way to hit targets that rewarded investors at the expense of organic market growth.
Buy in company after company in lip service to "strategy" you can sell to keep the Street happy. The internal focus on engineering excellence? What's that got to do with global markets?
It works for a bit, and then your pants fall down. Nobody worries about finding reliable hardware these days – they worry about executing complex projects in a complex data ecosystem in the service of modern business needs. Engineering excellence can make this happen. Nothing else can. It's a different sort of engineering to that which forges mainframes from 1970s technology, but it follows the same principles of analysis, knowledge, imagination and discipline.
- We've been shown time and again that strong encryption puts crims behind bars, so why do politicos hate it?
IBM can no longer pretend to excellence here. It has seemingly muffed a process that is the minimum viable product to sell to big business. Its self-evisceration of core competencies is such that it can seemingly no longer even support itself on its own terms for the most basic and well-established of business IT functions. And that's not a good sales talking point.
IBM has one chance of salvation: to migrate back to engineering. It must analyse its own systemic problems, plan a path of reform with imagination and pragmatism, and execute with discipline. And it must be seen to do it, or the dogs of cloud will tear it apart just like the clone dogs ripped out its desktops. But this time, there will be nowhere else to go. ®