As Technology Professionals, we have an obligation…
It’s pretty common for software projects to fail. According to the Standish Group’s 2009 CHAOS report, only 32% of software projects succeeded over the past year. Not exactly a banner year to be sure; but even in a good year, it’s rare to hit more than 40% success. Whatever the reasons and wherever the fault may lie, these are disturbing numbers that I think we, as technology professionals, need to take ownership of and do our part to correct. (For a little background reading, have a look at this excellent IEEE paper about software failures).
It’s easy for a member of a software project team – especially developers – to up and bail out of a ‘disaster’ project situation and simply move on to another project. There are plenty of opportunities out there for skilled technical people, so it doesn’t take much to just throw your hands up in the air and abandon the disaster for a more appealing career move. It’s equally easy to ride-out the disaster and just show-up for work, and tune out the world around you.
The thinking, of course, is that it’s someone else’s fault, someone else’s problem and someone else’s financial catastrophe to sort out. After all, you’re just a small player in a big machine. It was probably a lack of requirements, poor leadership, lack of focus, too many changes, management issues, etc etc. that caused the failure, right? So, why should you take on any responsibility?
The trouble with that type of thinking is, Those People left to pick up the pieces of a software failure are left with a bitterly sour taste about the risk associated with investing in software. They have to stand in front of investors, customers, regulatory bodies and other stakeholders to explain the multi-million dollar failure. And you know what? They are not happy with the 32% success rate in software projects. It’s both expensive and it looks bad on their career.
And with all these software failures on the books, They may decide that, when faced with the prospect of investing in software products & projects – and technology in general – it is far too risky and their money is better invested elsewhere. We really need to improve things.
So, without descending into a blame and fault discussion, let’s be honest about the burden of responsibility of software project success and failure. While it’s true that the causes are rooted in many influences and situations in an organization, my goal in this discussion is to zero in on what we, as technology professionals, HAVE TO DO to avoid and repair the ongoing brutal legacy of software project failures.
On the positive side, I truly believe that we’re getting better at building software. This is echoed in historical CHAOS reports from the Standish group, but it’s also something that’s evident in the many improvements and maturity in the world of software development: better tools, improved standards, and the emergence and more widespread adoption of Agile and accompanied development processes that reduce risk.
Part of the challenge is that there really aren’t any established standards in this industry. Adopting standards is optional. There are obviously scores of strategies to improve quality and the ability to reliably hit release targets, but it’s typically at the discretion of the development team and mostly overlooked … which is largely why we’re at a 32% success rate.
So, to start with: whether you’re a manager, developer, product owner or business person, here’s a bare-minimum list of what your software development team should be doing to improve your ability to build and ship quality software:
- Continuous build on an independent build machine. Check out Continuous Integration for more.
- Improve team communication through daily Stand-ups or Scrums
- Integrate QA into your development team and development cycle
- Keep your software in a continuous working state. Only the code that’s checked-in should be considered as valid working code.
- Ship your software product often – like every week. Even if it’s just to internal stakeholders.
- Divide the release cycle into manageable chunks (sprints, iterations)
- Remove barriers between the technology teams and the rest of the company or to customers. Invite users/stakeholders to regularly see your working software in progress
- Adopt best practices around: peer reviews, QA, exception handling, coding standards.
Each of these simple process improvements vastly reduce risk and increase the chances for success. These examples are only a scratching of the types of methodology/process upgrading that teams should be adopting, but I believe if ‘most’ teams were to do only those things, the number would dramatically move from less than 40% to well over 50% success. Still not great – and I’ll address some more deep reaching improvement concepts in upcoming discussions – but if more teams and technology professionals were to adopt a sense of ownership in improving the global perspective of software project reliability, we can avoid a most certain backlash in the availability of technology investment dollars.