Michael Nygard: We need to take a closer look at how IT is failing
In his opening keynote speech to the W-Jax 2014 in Munich, Nygard told us what’s going wrong in the IT world and what needs changing.
It seems like it’s never been a better time to be a developer. Most of Silicon Valley believes that if you want success, you need to program it yourself. But the reality behind IT’s success-driven facade is not so rosy.
Two thirds of IT projects fail, says Michael Nygard, best known for his acclaimed book Release It! At the same time, 80% of IT budget is spent on maintenance.
We fail. A lot.
This year’s opening keynote speaker at the W-Jax gives us the example of a failed attempt to modernise the IRS, which sent $4 billion in government IT budget down the drain. “And not one light of that code ever saw the light of day,” Nygard tells us. The US government has already invested $8 billion into a new phase of redeveloping the US revenue system – it’s still not ready.
Nygard is making an important point here: In spite of IT’s major success stories, we need to take a harder look at why we fail so often.
Risk management does not mean risk reduction.
IT companies have developed numerous practices to reduce potential risk: matrix management, standardisation, offshoring and so on. But Nygard argues that these very risk-avoiding procedures can also introduce new kinds of risks.
By pushing us towards monolithic development, these risk management procedures make our systems harder to control, and by extension, expensive. And that’s what risk management wants to avoid, right?
Our users are smarter.
We’re now beginning to see a second generation of web natives entering the workforce. “I think the last president that has his secretary printing out emails will retire in this decade,” Nygard jokes.
But is this making it harder or easier for IT? While a smarter audience makes it possible to develop more complex user interfaces, the expectations of our users has never been higher.
Waiting for approval sucks.
Some of us probably already know how annoying IT approval processes can be. But Nygard believes that it’s more than just a nuisance, it’s one of the biggest contributing factors to delays – which again means cost. Projects can often be forced to stop and start because your approval request is waiting in somebody’s inbox.
Sometimes there are forms and meetings and powerpoint presentations involved. It might just be a few hours, but this waiting time is what over-budgeted, delayed IT projects are made of.
Decentralization isn’t the only way forward.
Decentralization has become one of the central aspects of modern IT. But when users need four passwords to interact with various departments (or systems) of one company, you have to wonder if decentralization is always the best road to take. Nygard argues that rather than switching from centralized systems to the entire opposite, the ideal system will be based on a mixture of centralized and decentralized, that acts like a network.
Developers need more freedom.
Nygard tells us a story of a system in an unnamed company that needed to be updated. The company was advised that it would take six developers six months to complete the project at a cost of one million dollars. Two employees decided to have a go at it as a hobby weekend project. The had it up and running by the end of the following week.
“If you free people up to do the best thing they can, they usually do the best thing for your business,” Nygard concludes. “We all know the feeling of relief you get at five o’clock when you know you can now start to get some actual work done.”
At the same time, Nygard argues that we need to allow application development and operations to make central business decisions, rather than preserving them in IT. “It’s time to undo that goldfish bowl and get people back in the business.”