olly - Fotolia
IT project failure isn't completely pervasive, but it still happens far too often. One 2017 survey reported that only 34% of projects finish on time and only 42% come in on budget. A Wrike study from 2015 reported a better outcome: 64% of projects meet their goals. But 70% of the companies polled reported having at least one project failure in the last year.
Granted, these numbers represent just two data points. But let's say those numbers are an exaggeration, and the reality is that the success rate is better than these reports indicated. Let's say your success rate is 80% -- even that is much too low when you consider that one failure can cost millions.
Can you imagine if the construction industry posted 20% failure rates year after year? We'd still be living in grass huts and igloos.
Yet, the world is hungry for software, even software that sort of works. Remember when Microsoft sold 20 million copies of Windows Vista in the first month of release? We all remember what a great OS that was. Car salesmen probably salivate at the idea of a buying public with such low expectations. Luckily, manufacturers can't let a car that sort of works roll onto the highway. Yet, creators of free and open source software can throw anything out there and let the market figure it out. As the MIT license states in uppercase letters, "THE SOFTWARE IS PROVIDED 'AS IS,' WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED..."
You get the message. A successful IT project is more the result of a favorable roll of the dice than a predictable, scientific endeavor. There are some strategies, though, that organizations can employ to give themselves better odds at avoiding IT project failure.
The importance of an accomplished project manager
The specter of IT project failure casts a dark shadow. No one wants to work for weeks, months or maybe years only to have the plug pulled before your code sees the light of day.
The source of these failures has been well-documented: incomplete requirements, unreliable estimates and/or technical incompetence. What's amazing is that, year after year, organizations face the same problems. The only difference between the project that succeeds and the one that fails is that failures suffer from scope creep, poor planning and completely inadequate resources and expertise. These issues will always be there, and that's why project managers are so important. Good project managers know how to get the dice to roll their way.
Organizations need project managers with proven track records, along with the technical, analytical and people skills -- as well as the chutzpah -- to guide a project to success. And, if the company is big enough, it can fund a project management office staffed by experienced PMs, which will increase the odds of success. It makes sense. A good field manager increases the chances of a baseball team getting to the playoffs, provided the given team has the necessary talent. The same can be said of IT projects.
Methodology doesn't matter
Now, it's time to consider development methodologies. Do the odds of success increase with Agile vs. waterfall? Not really.
A Standish Group study revealed that, under Agile, 42% of the projects surveyed were delivered on time, on budget and with all features expected. Under Waterfall, 26% of the projects met the same criteria.
Go with Agile, right? The numbers indicate a better chance of success. Well, not really. When you look at the percentage of projects that were deemed challenged -- projects that were delivered but failed to meet schedule, budget and/or features goals -- Agile comes in at 50%, while Waterfall is 53%. This means that, given the choice between Agile and Waterfall, the probability of not achieving full success is about the same.
Small teams have a better chance of success
There is one statistic, though, that indicates a high rate of project success. A 2013 Standish Group report found that projects with a payroll of less than $1 million have a success rate of 76%. Projects with a payroll exceeding $10 million, meanwhile, have a success rate of just 10%.
Let's apply this to the real world. A project with a payroll of $1 million translates roughly into a team of around 10 employees. A project with a payroll of over $10 million might have 10 times the number of project managers, developers, test practitioners, release managers and project assistants. As we add more people, the complexity of the technology and personnel administration increases, as does the number of cracks through which things can fall.
If there's one thing you can do to head off IT project failures, it's this: Keep the project team to under 10 people.
You might say that a small team is great for an e-commerce startup, but what do we do when it comes to a large project the size of HealthCare.gov?
My advice? Do the same things we do to keep something as big as the internet running. Have small teams aggregate into a larger whole by promoting interoperability and collaboration using open standards, open frameworks, open reporting and open source development methodologies. Is it easy? No. The details make all the difference.
History has shown that successful IT projects are more the outcome of an informed game of chance than a reliable practice of science. Thus, the trick is to go with the odds. Small, well-funded projects staffed by competent professionals under the supervision of an experienced project manager have the best chance of success.