What stops a BI implementation from being a success?

Now that is a good interview question - the simplicity of the question hides a vast richness of answers; it can differentiate between the "been there, done it" and the "read a book"; it can tease out the roundedness of a candidate - are they reporting tool biased or database centric. It can also separate the people who actually think about the question in a larger picture than just "what went wrong with my last implementation".

Now how to you measure success? A reporting tool vendor may measure success in the market by licenses sold (a database vendor may similarly measure number of systems or terabytes managed) but how do the number of BI seats equate to satisfied users? The answer is of course they don't. Success comes from satisfying needs, and for BI that need is information. Simplistically, this boils down to three criteria: getting the information in a timely fashion, easy (and complete) access  to the information required, and trust in the quality of the information. Get any of these three wrong and then the user base will smell a failed BI implementation, regardless of whether the functional and non-functional requirements for the system are met.