1) Get the Numbers: The best way to avoid setting unrealistic expectations for future projects is to get data from completed projects. Metrics should include timing, cost, quality, and size and complexity. If you start collecting this data on current projects, over time you’ll be able to do a “before and after” analysis.
2) See Software Interdependencies: We tend to forget that software development is not linear, but geometric. Double the amount of gas in your car’s tank, and you can drive twice as far. Double the size of a software project, and the number of data points can easily grow four- or even sixfold. Make sure your project plans take this into account when you add features, requirements, or resources to an IT project.
3) Know Talk Isn’t Cheap: Appreciate the complexity of human communications in an IT environment. Techies call this the signal-to-noise ratio. To keep the noise down, break all projects into small pieces that can be handled by teams of just three to five people, especially when shipping work offshore. If you have programmers in India being managed from New Jersey, build in time and resources for these two groups to communicate effectively.
4) Benchmark Project Performance: IT analyst firms are great at creating macro, industrywide data. But what you need is project data. For example, what’s the average amount of time needed to install the kind of payroll system you’re thinking of? This data is available from third parties, often at minimal cost.
5) Walk the Talk: Remember that project workers will fear being measured until you can show that the data is being used to improve the company’s productivity. You must have senior-level people driving the measuring process: project-level people already feel they’re the victims of unreasonable expectations. Show them that data will be aggregated to identify systemic problems — not just to punish individuals.
6) Monitor Performance — Quickly: Track your total IT portfolio in real time, not just quarterly or monthly. You want to know if the project is running over budget or late, of course, but you also want to know how it’s faring on reliability, functionality, and complexity. Essentially, you want to see a yellow light while there’s still time to make changes.
IT project success rates have doubled during the past eight years.
Source: The Standish Group International
- 1994: 31%
- 2002: 15%
- 1994: 53%
- 2002: 51%
- 1994: 16%
- 2002: 34%