Do users/customers know best?

This has been one that's puzzled me for a long time. I remember a JN Alertbox ("Are Users Stupid" http://www.useit.com/alertbox/20010204.html and "Don't listen to users" http://www.useit.com/alertbox/20010805.html ) that warned against accepting everything a user suggested and I'm mindful of much of the management-world's concern around suggestion boxes - principally that they generate an awful lot of noise and resentment for very little innovative gain. In essence this boils down to a human trait that we (generally speaking) don't like to hear that someone knows better, that we haven't acknowledged a problem and that, when we believe we're right, that we struggle to accept alternative viewpoints. As usability practitioners and advocates of a customer-centric philosophy are we biting off more than we can chew by attempting to please all of the people all of the time? How do we strike the balance between integrating user-feedback with our own objective (oft empirical) understanding of what works and what doesn't? Customers and users seem to be more able at telling us what they don't like in preference to what they do like, and consequently it's much easier to eliminate and reduce than it is to innovate. When customers do tell us what they want we can often find ourselves (metaphorically) furrowing our brows with the thought "I really don't think that's going to work". Should we be giving in to this thought and applying our own usability spin on it or should we accept the users comments un-appended and without question?

I suppose the answer lies in that old chestnut, 'a bit of both'. Whilst listening to users and customers is useful, actually observing how they behave is absolutely critical. Someone may say they find a site easy to navigate because they are satisfied in the expectation they will need to explore but you may observe them taking a significant amount of time to perform basic tasks. Implication: Make the navigation process more straightforward and you take the user from being satisfied to being impressed. Users may well express satisfaction with a site simply because it looks good (what is beautiful is usable), before they make an assessment on usability. They may over-rationalise their behaviour, assume mistakes are their own ("silly me, I didn't see that link") and play to what they think you want to hear. Is a customer invited to a focus group and plied with refreshments really going to recall their true experience?

All this has come to a head in my mind because we're talking about adopting a suggestion-box scheme (albeit an intranet-based one) here at work to capture employees thoughts about improving the customer experience. I remain cynical. I still think that there's a need for leaders to be leaders and for experts to be expert. The effect of sorting and evaluating the most credible suggestions will be highly resource-intensive, it'll be subjective, it'll create resentment in those whose ideas were rejected and there will be an exponential drop-off in submissions as people feel disheartened that their idea was not taken up. There are ways, of course, to combat this through constructive feedback on suggestions, championing of successful suggestions and other incentivisation but the potential for noise, in my mind, outweighs the potential benefit. The role of consultants in this case must be to see the wider picture, can front-line staff see the processes behind their experiences - do they need to see them to know whether or not they're broken or indeed changeable?

Ultimately I'm starting to feel that pure customer-centricity or user-centricity is not enough and that taking second-person (user) perspective is worthless. The hardest challenge for me will be to ensure that, when in third person perspective I given enough credence to users and metrics alike, performing the role of mediator to ensure that, ultimately the user/customer-experience is the winner.