Failure is always an option – part 1 (chaos)

One of the most popular reports people use to showcase failure of software development is the chaos report from The Standish Group. The Standish Group collects information on project failures in the software development industry in an attempt to assess the state of the industry. 

In 1994, they reported a shocking 16 percent project success rate, another 53 percent of the projects were challenged (not on time, over budget and with fewer functions than originally specified), and 31 percent failed outright.  Although the newer reports show better numbers, the overall results still paint a dire picture:

There aren’t a whole lot of other statistics out there on this topic, so obviously these numbers get big play. Guilty as charged, your honor. I have used them myself, in a presentation or two.

I won’t be doing that again.

I realized that I have some serious problems with these metrics. They measure a project’s success by solely looking at whether the projects were completed on time, on budget and with required features and functions. But what they do not take into account are things like quality, risk and customer satisfaction. Could it be that an extremely unstable, unusable and frustrating piece of sofware that was delivered on time and on budget qualifies as a success? I beg to differ.

The Standish Group’s methods are not fully disclosed, and the bits that are disclosed are apparently deeply flawed. Their figures are misleading, one-sided and meaningless – the results are completely unreliable. They present their figures as absolute facts, but I lack clear context. The most famous sceptics of the report are Jørgensen and Moløkken. They emphasize its unreliability and question the claim of a “software crisis”:

” Even the definition of challenged projects is not easy to interpret. It is defined as “The project is completed and operational but over budget, over the time estimated, and offers fewer features and functions than originally specified.” The problem here is the use of “and” instead of “or”, combined with the following definition of successful projects: “The project is completed on-time and on-budget, with all features and functions as initially specified.” Consider a project that is on-time, and onbudget, but not with all specified functionality. Is this project to be categorized as challenged or successful? Our guess is that it would be categorized as challenged, but this is not consistent with the provided definition of challenged projects. ” 

In the comments section of an interview with The Standish Group’s Jim Johnson, Jørgensen brought up his critique of the CHAOS report and asked Johnson two very fair questions. Johnson’s reply is pretty enlightening, to say the least. Here are a few excerpts:

…We are an advisory research firm much like a Gartner or Forrester. Neither they nor we can afford to give our opinions away for free. We have facilities, utilities, and personnel and we must, the same as you, be able to pay our bills. Just because someone asks a question, does not mean we will respond with an answer. In fact, we most likely will not…

…Our current standard answer to a CHAOS inquiry is, first: please purchase our new book, ”My Life is Failure” in our online store. If that does not satisfy you, then you need to join CHAOS University. If you do not find your answer or answers there then you need to purchase our inquiry services. Then we will work to answer your questions…

…It is strange that Jørgensen has never applied or professed interest in joining us. Some answers can be found if you join us at CHAOS University 2007 or one of the many outreach events. So you can contribute to the CHAOS research by providing funding or sweat, but short of that you will and must be ignored by design…

Don’t get me wrong. I think there *are* lots of failing software development projects, but in other numbers and for other reasons than the ones Standish brings forth: deliveries that do not bring any value to its users, software that was poorly tested or poorly designed, resulting in failures in production.

The problem I have with the Chaos Report is that they claim to be some kind of “industry standard”, projecting a false image of the dire state of the software industry, based on poor metrics. And I certainly don’t believe in the “quality is dead” mantra that resonates from their reports. Sure, there’s plenty of chaos out there, but I like what Henry Miller said about that : “Chaos is the score upon which reality is written”.

I’m with Henry on this one.