[tawgx]

I think the best way to work with customers (and I do believe you should) is not having theoretical discussions with them about the market or whether a new product can take off, but rather give them a number of tangible alternatives (can be mock ups) to play with and see how they react to them.

[benjaz]

Often, customers don't see the consequences of "small" changes that they'd like to see. At AOL we had so many people giving feedback on products that it became a customer support issue. One solution proposed was to have a dummy site that allowed users to move elements around. Whenever they moved something on the page (or dragged/dropped from a dynamic list of popular suggestions) they'd see the consequences in real-time. Big bold messages declaring all the changes that were just made would make it obvious that they need to think through their ideas.

[epo]

If you are devising a new product concept, the very first telephone (or ipod or ipad) say, then what is gained by doing this? Your customers have no point of reference. In general, asking customers what they think is only useful if you are iterating on the known and familiar. Otherwise it is just harvesting random opinions, most of which will be variants of "make it more like what we already know".

[Ciotti]

So mostly relying on reactive feedback rather than proactive suggestions? I agree.