The issue here is that it cleaves toward designing for the audience you have, as opposed to the audience you seek to appease.
Often, the design goals might exist on an alternate axis.
Random people offering random opinions amounts to random noise. However, the random noise will not appear neutral; it will appear as information.
If we consider many different things designed for particular audience members such as jet cockpits, medical tools, racing automobiles, we will see traits that exist that may seem nonsensical or otherwise when we divorce them from their designed contexts.
Bill Buxton covers this in Sketching User Interfaces when he describes Inuit coastal maps:
"The Inuit have used. [...] Tactile maps of the coastline, carved out of wood. They can be carried inside your mittens, so your hands stay warm. They have infinite battery life, and can be read, even in the six months of the year that it is dark. And, if they are accidentally dropped into the water, they float. What you and I might see as a stick, for the Inuit can be an elegant design solution that is appropriate for their particular environment."[1]
Focusing on complaints of the above design in all likelihood would, given the mad rabble of audiences online, result in discarding a solid bit of design.
[1] Bill Buxton, Sketching User Experiences, pg 37
Listening do user complaints doesn't always mean listening to their suggestions in fixing the problem. You shouldn't necessarily remove something that every user is asking you to remove if you feel it's a critical feature, but it definitely means you need to rethink how the feature works.
Since that change, I haven't heard word one about our terrible, onerous, awful default body and title character limit policies. Not one. Single. Complaint.
So, they didn't do what the customers said they wanted, but they did work on the issue until people quit complaining.
“If I had asked people what they wanted, they would have said faster horses.”
― Henry Ford
Plus, IIRC, for the Edsel, they did ask customers what they wanted. And it has gone down in history as one of the worst cars ever.
Yeah. Jeff didn't do what the customers wanted and he's smart enough to not let users sway the design. I think the issue here is that it is a potential pitfall that should have garnered a sentence or two for less experienced people that read the article.
Isn't it more likely that the tactile maps are used because they are superior to any other kind of map the Inuits could have created (non tactile drawing, or a rock that doesn't float, for example)?
A tad of a chicken-egg dilemma there. While I can only speculate, I'd surmise that the Inuit are / were well aware of their contextual circumstance and designed accordingly, and quite likely iterated on the design.[1]
The issue I raise is a cautionary tale that when we try to use some abstract evaluation well removed from the context, the "data" can end up being non-information. Worse is when design evaluation comes with hidden cognitive bias or underlying ideology.
That is, even a valid concern / complaint of a given design my have no merit given the design context. If we take for example a vehicle, one might complain that it has a lack of cargo space or a child seat restraint. Valid complaint given a very discrete set of design contexts. To follow along the original example in the post, these otherwise valid complaints of a design are moot if the vehicle in question is an F1 automobile designed to be raced in unique circumstances with unique needs.
[1] There is also the possibility that some of the design contexts include more abstract needs such as cultural, religious, historical, aesthetic, or even the darker notions of power dynamics. To diminish the potential influence of these facets would be an ignorance in our design evaluation.
Often, the design goals might exist on an alternate axis.
Random people offering random opinions amounts to random noise. However, the random noise will not appear neutral; it will appear as information.
If we consider many different things designed for particular audience members such as jet cockpits, medical tools, racing automobiles, we will see traits that exist that may seem nonsensical or otherwise when we divorce them from their designed contexts.
Bill Buxton covers this in Sketching User Interfaces when he describes Inuit coastal maps: "The Inuit have used. [...] Tactile maps of the coastline, carved out of wood. They can be carried inside your mittens, so your hands stay warm. They have infinite battery life, and can be read, even in the six months of the year that it is dark. And, if they are accidentally dropped into the water, they float. What you and I might see as a stick, for the Inuit can be an elegant design solution that is appropriate for their particular environment."[1]
Focusing on complaints of the above design in all likelihood would, given the mad rabble of audiences online, result in discarding a solid bit of design.
[1] Bill Buxton, Sketching User Experiences, pg 37