The truth about software reliability

Research results: It’s not just about quality

Whether it’s systems-level platforms and tools or full-blown business applications, you want the software your organisation relies on to run reliably and predictably. Our recent Reg survey, however, confirmed what most probably knew from experience, or could guess from press headlines, i.e. that software-related failures are far from uncommon.

Unsurprisingly, major system failures and/or incidents leading to loss of data were top of the list when it came to the potential for damage and disruption. What also came through, though, was how much people hate it when a piece of software inflicts a constant stream of minor issues on users. Some likened this to ‘death by a thousand cuts’, with satisfaction and confidence ultimately completely undermined even though no single incident was that severe.

Such findings highlight the importance of software quality management on the part of suppliers, and the evidence suggests that many have work to do in this area. A significant number of suppliers, for example, confess to knowingly shipping software that has failed internal tests because the problems flagged up couldn’t be diagnosed and corrected in time.

Troubleshooting and diagnostics was, in fact, a recurring theme in the study, especially concerning complex production systems. With myriad dependencies and influencing factors, the consensus was that tracking down the root cause of software issues is a constant challenge. Troubleshooting intermittent failures was a particular problem highlighted by many.

But how you respond when software failures occur matters a lot. One of the biggest frustrations is when problems drag on, while no one can explain the cause. A diagnosis and plan to remediate can buy you time before satisfaction issues occur, but troubleshoot and fix things quickly and you could even achieve hero status.

Perhaps the main conclusion from the study was the need for software suppliers and the in-house IT team to cooperate more freely and work together more proactively to resolve significant problems.

Respondents from the customer camp highlighted that some suppliers habitually do their best to avoid taking ownership and responsibility. Stating that the software was never designed for that kind of use in that kind of environment doesn’t help when users are screaming at you for a fix. Suppliers, meanwhile, talk about customers sometimes not helping themselves. Inadequate training and failure to disclose relevant information during the troubleshooting process are examples here.

Against this background, the research, which was sponsored by Undo, also touched on other areas, ranging from the need for better account management on the part of suppliers to the way in which advanced diagnostic tools can help. If you want to know more, you can download your copy of the report here.

Originally published on The Register