Here's another symptom of a Request for Faerie Dust™:

A proposal that includes (1) analysis, (2) proof-of-concept, and (3) a plan for a solution.

Sounds sensible, doesn't it? Study it, prove that you can solve it, and write the complete plan to solve it.

Sounded sensible to me, until I had to write the plan.

Let's say we spend 3 weeks meeting the users, figuring out what's wrong, brainstorming root causes. What, then, will the problem turn out to be? And what, precisely are we supposed to propose now before having actually done the analysis?

Before we do the analysis, we can't rationally plan a POC or even set a budget for a POC. Indeed, we can't even be sure that a POC is needed.

Unless -- and this is my thesis -- the problem is already firmly in someone's mind, and they want me to sprinkle Faerie Dust on the known problem to make it go away.

I suspect that any change that's outside the box will be unacceptable. For example, it is most likely that their training is inadequate, and some of their content management practices aren't quite right. Their staff can't find things in the knowledge repository because the titles, names or index keywords aren't what people think they should be. However, they want Faerie Dust sprinkled on Search to make it "more effective".

Clearly, they expect the analysis to turn up problems with search. The RFP listed a bunch of candidate suggestions for improvements, most of which were performance-related -- i.e., speed up search.

The good news is that some of the candidate suggestions for improvements were organizational. In this case, the POC isn't purely technical, and will require scheduling people like information designers, authors, trainers or users. How do we propose that work?

And how do we know -- without doing the analysis -- what we are proposing for the POC?