I received an intriguing email this week asking this question: ” [..]if we were to put a quantitative value against each best practice, summed them up, and compared the total against a possible maximum could we have a predictor of project success?”
No is the short answer. Here’s why: People need to first select which practices are best for their project, and then assess the practices to see if they are working for the project. Not all “best” practices are best for each project. I’ve worked on projects where daily builds were too infrequent and projects where hourly builds were too frequent. If people aren’t trained in inspection, then inspections aren’t appropriate, but peer review might be. I see many projects where the requirements would be better served by not writing anything down except the location of the white board where the real requirements are (and then transcribing those onto some less ephemeral medium.)
As an example, I once consulted to an organization that dutifully recorded its defects discovered per inspection, but did not record prep time. People stopped preparing, and it looked as if the inspections were going well, but the number of defects found kept dropping. Management convinced themselves they were following a best practice of inspection, but the product was so bad it died on introduction.
If you want to predict project success, first determine which practices make sense for the project. Then determine what you need to qualitatively and quantitatively measure about each practice. Make sure you think about all dimensions of the practice (remember my six sides of the project) to make sure the measurement is not single-dimensional.
In my experience, the best predictor of project success is the people on the project. If they know how to work together and they are capable of performing the work, the work will complete with as much success as the environment allows.