Are you able to get your projects done early or on time?
Yes – 67%
No – 33%
Software Development Times – Data Watch, Nov 1, 2010, Source: Evans Data
This report in Software Development Times caught my eye. When talking to software developers, IT managers and project managers in general, I usually get around to asking about how often they delivered their projects on time. The answers I get are pretty consistent.
They generally start off by saying they do OK. They deliver projects, or at least their part of the project, regularly on time or even early. I then ask about how the overall project went, how many days or weeks it was early or late, and did they ever have to replan a project?
Organizations that deliver on time have a tendency to know their numbers, such as how long projects took and how often they delivered on time. Those that say things are going pretty well but ultimately I find out that they are not, have a tendency to speak in very general and expansive terms. The one I always like is, after getting the positive sweeping generality that they deliver on time and I then ask some probing questions, I will get a suspicious pause and then the question “well, what do you mean by on time?” (For my answer to this question see Is Your Project On Time?).
The numbers reported in the Software Development Times were for open source developers working on open source projects while working their salaried job (so admittedly a unique bunch of people). While this may be a different breed of people, the number of 67% delivering their projects on time or early raised my “I’d like to dig deeper” senses.
As I related in other articles (see Knowing Your Project Management Average), even people who are well respected for their expertise and performance usually have a very hard time remembering how long things took to complete, what was worked on, and how well it went. I’ve read similar findings to these — reported at least as far back as the work of Peter Drucker in the 1960s. Once we measured actual task completion times, for example, the experts consistently underestimated how long the task had taken by two to four times even when they were directly involved with the work.
While I only know about these numbers by what was published in the Software Development Times, they sounded a bit optimistic based upon my experience and research. Experience also tells me that this kind of data is most accurate when coming from a reliable record of what really happened, and not exclusively from the memory of otherwise very competent but extremely busy and hard working people.