Re: The Mythical 5%
Posted: Jan 3, 2008 11:17 AM
> Bruce writes good books and I respect his leadership in
> the software engineering profession, however, this
> "software crisis" myth is getting old. It's an overused
> cliché that all well know software pundits over use. I
> recently heard on by Douglas Crockford, he starts the talk
> with the same "sobering" (say it with a quiver in your
> voice) statistics.
I guess it depends how you define "fail". In big companies, I would guess (just from what I've seen across many companies) that over half the projects are still-born (i.e. started but never go all the way to end user production use).
If you define "failure" more loosely, as in "funded projects that make it past the initial 90 days (first trimester) and get finished enough that they could theoretically be deployed", then the failure rate is probably 10% or 15%.
On the other hand, if you define "failure" more strictly, as in (a) significantly over budget, (b) significantly late, (c) failing to achieve major requirements, (d) failing to show return on investment, (e) failing to meet a reasonable quality level, and (f) lacking the ability to be cost-effective maintained or survive the departure of key personnel (i.e. risk mitigation), then I would guess that 95% or more of software projects "fail".
Cameron Purdy | Oracle