Home | Login | Recent Changes | Search | All Pages | Help

ProjectFailureRate

In any serious discussion about the terrible state of practice in software development, chances are good that someone will bring up the now-famous 1994 Standish "CHAOS" Report, which reported that 31% of projects in their survery were cancelled, and only 16% of projects "succeeded" (the remainder being over budget/behind schedule).

Someone is finally trying to take a serious look at how Standish came up with those numbers.

I'm guessing that the Standish methodology won't stand up to scrutiny, but that their numbers won't be too far off the mark, if it's even possible at this point to back up 12+ years to set a mark.

The satisfying part for me is seeing someone take a serious look at the history we've been fed.

--DaveSmith 2006.08.03

"Project success and failure" might have something to do with effectiveness so we can kick it around ourselves if we want to. I've been mulling that Standish report myself for some time - about ten years actually - and had more than a few questions about their methodology from day one. Some of us are skeptics to begin with. The CHAOS report was, and is interesting fodder. (Before that it was the "data" trotted out by the SEI and PMI. Somewhat later than the CHAOS report it's the tales of how bad and awful things are from the big-A agilist folk. Every day it's both the problems and solutions I think I have.)

I become inspired to look at that report pretty much every time some solution peddler trots it forward to establish the problem they propose to solve. When some solution seller tells me I have a problem I didn't know about before, I hide my wallet first. After that, I may listen to them further.

- JimBullock (I am somewhat of an expert on failure myself. Should anyone listen to me about how to succeed?)

What if I'm an expert on other people's failures?

- Anon

I wonder. What kind(s) of evidence *do* we have? What kind(s) of evidence would be convincing if we had it?

Anybody?

- JimBullock (It ain't what we don't know that gets us into trouble. It's what we know that ain't so. - ?Will Rogers? Or am I wrong about that?)

This project success and failure topic has bothered me much in the past few years. From 1998 through 2001 I was involved in a project (the government project manager) that cost $33M instead of $18M and was 6 years long instead of 5 years. I believe that project was a success. It gave the users a major capability that they needed and did not have. Since delivery, the systems have been in constant use in the field.

My current project is over budget by 15%. That is 15% more than what people estimated before we really knew the requirements. I think that is a success. I have sensior managers, however, who feel this is a catastrophic failure.

DwaynePhillips 8 August 2006

(P. S. I just noticed that if add the digits in each date field 8-8-2006 you get 8-8-8, which goes to prove a theory of one of my colleagues that you can do just about anything with the numbers in dates)

Now, I wonder, how many different ways could that same, single observation be "a success" or "a failure?" That's an interesting question. - JimBullock

If you do a feasibility study and decide not to go forward, is that a success or a failure? Doesn't it depend on what risks you avoided?

If you build something, and put the first increment into operation, is that one project? Is the second iteration a second project? What if you have several functional increments going at the same time, each with its own team? How many projects is that?

There are so many questions of this type that I think "project failure rate" is a meaningless idea. Then, to say that 31% of projects fail--well, that 1 after the 3 is absolute proof to me that this study was meaningless, and I don't have to read another word to know it. Come on, folks, do you really think that anyone can define project, success, or failure to two meaningful digits? - JerryWeinberg 2006.08.13