Nice post.  The technology does often seem to be an easy scapegoat for larger issues.  In fact, I would argue that one of the main reasons to even go through a CMS selection process is to refine the requirements/need for even using a CMS (and trying to prioritize and not just throw in a bunch of buzzwords or laundry list).  Of course, another broader issue of CMS success is whether the organization has strong Web Operations Management (welchmanpierpoint.com/.../wom-primer).

Thanks for information David.   I agree about the requirements side.  It is a big focus for us when helping a customer to find a CMS  We talk about it in terms of 'Fit Factors' that cut across a lot of areas.

All excellent advice. I would only add:

Something I hear all the time from organisations: "I think the CMS is fine, but our implementation was poor". This is well and good, but when I hear this for *every one* of the sites from some products, you have to question the CMS.

So I would advise checking around: if no-one else has had success with the CMS, this may point to a new product...

I've had a few ideas in the past in a similar vein:

contentedmanagement.net/.../is-it-what-or-how-that-broke-it

I disagree with James Robertson though. Most of the time I hear it's the technology that's poor, when actually people just don't know how to exploit it.

James and Phillppe you actually are both right in a sense.  You have to remember every CMS was originally created to 'scratch some itch' and then adapted over time to try to serve everyone under the sun (or get every checkbox on CMS Matrix).  That simple fact alone predisposes each CMS platform to be stronger in some scenarios and almost useless in others.

A big part of our jobs as CMS professionals is helping people see past the shiny demo and understand how the CMS works specifically for the customer and their key scenarios. (Please queue Matchmaker from Fiddler on the Roof).  

Now don't get me wrong there are some CMS platforms out there that are more marketing hype and slick demos than tangible product. Yes you can build a site on on them, and yes it will work but in truth they kind of suck and should be replaced.

So to wrap it up you have to also make sure the CMS matches intended site usage/scenario as well as the customer's specific requirements and make sure the product viable (and about 1000 other things that are not specific to this post).

The habit of blaming the technology is not restricted to CMS. I've seen the same processes occur around search tools and other applications.

The challenge is lifting the debate from the technology (tool) to the desired outcome s and the necessary processes used to achieve them.

If discussions can be refocused around outcomes people can often begin identifying the real issues, which are often very addressable within the technologies in place, or can be added with a third-party add-on, rather than by throwing the baby out with the bathwater.

Interesting article.  I'm approaching this from the perspective of a solution provider  implementing the site using a customer selected CMS product.  The reality is that when we are engaged by the customer, the cat is already among the pigeons.  The IA is a works in progress, the designs incomplete,  the content model wobbling like a jelly, the content strategy is vapourware and the due date looming.  It's my personal recurring nightmare.  

From your list, I think planning and the lack of strategy are real killers.  Changing requirements are the norm and a poor technical implementation works if it meets that requirements.  So third the prize would be training.  Always vastly underrated...