The list of unusable products is long, here I try to focus on those that were made viable by users.

Boeing airliners:
The 727 and 737 were not nearly reliable enough on entry into service. They became adequate or better throught the efforts of both Boeing and its customers. In many cases customers led the way in designing and organizing.

Flo n' go fuel pump:
A modest hand-squeeze pump to get fuel from a container into a lawnmower or other small engine cleanly and conveniently.
But this device, patented no less!, has a handle latch that makes it unusable. With vigorous pumping the handle will travel far enough to trigger the latch, requiring release by your third hand.
So the only way it is usable by most people is to dissassemble it and remove the latch, , which doesn't seem to have much value as the handle is spring-loaded to the no-flow position in any case (needed to prevent siphoning).
While doing that, add some convoluted wiring-loom cover material to the hose near the handle to prevent kinking of it.

(And if you instead want to pour the fuel into a road vehicle using a container with spout, as you might if you run out of gas, beware that the current fad is various spout design schemes to prevent fuel spillage - but those of us with only two hands will spill more than with the traditional angled rigid spout, which is less costly to make. Duh?)

A powered jar opener:
Someone designed an appliance that tightens onto a jar and its lid then proceeds to unscrew the lid. Bulky, but just the thing for people with weak hands. (Those with strong hands can use simple manual devices.)
But the powered device requires an understanding of how to release the top down onto the jar from the raised position it gets into for removing the jar. Why not mold a simple instruction in the top?
Worse, the release latch does not always work. Why?
The base of the latch catches on the handle. Examination of internals shows that the parts go together in a way that cannot create much variation but do have a bit of slop so they don't bind together. There is either a basic design error, or an error in translating the design into molds for volume production. The problem is simply not enough clearance provided - so sometimes clearance goes negative.
The clever user will cut off a corner of the latch base, at an angle so that if the latch contacts the handle it will be moved aside by the new geometry of the two parts.

A medical appliance:
The ADAMS pillow system of nasal interface for Continuous Positive Airway Pressure to prevent Obstructive Sleep Apnea is the best nasal interface concept. But the headgear does not function well enough for it to be a viable interface.
Enter an entrepreneur who produces much better headgear specifically for that device.
And a user who made his own pillows to better seal against the nose - he looked at the basic anatomy of most people's noses.
Next is making his own manifold to replace the clunky original.


The bigger picture:
But what about all the users who aren't able to make the product work, or take the time to? How much do you lose in sales because retailers get too many returns or users tell 19 friends? (19,000 today with the Internet.)

And hmm - others now have the knowledge to make a complete set of parts to replace your CPAP interface product. Your patent has expired. Where does that leave you?

How do you avoid those traps?
Keith has suggestions.