Users don't come to you with problems, they come to you with shitty solutions.
You need to take their solution, reverse engineer their real problem from that, then work out a good solution for that problem that fits well with your product.
And I have talked with our guys and I can sort of see it from their view-point. from their point of view it often is the hunt for the last `checkbox`, the last necessary tick somewhere on the customers requirements page.
In their mind it often seems as a simple calculation, on one hand potentially so much $/year from big-corp vs 2xManxMonth to get a new feature over the line? No brainer.
An I am struggling to somehow explain that you need to add much more to those two developer months, because with every feature there are bugs and regressions and customer tickets, e.t.c. and suddenly it takes one developer in perpetuity.
A couple recent related articles:
1) How to Pick Winning Product Features https://medium.com/pm-insights/how-to-pick-winning-product-f...
2) Building Products https://medium.com/the-year-of-the-looking-glass/building-pr...
To be honest, it is much better now, but as a QE on the project, now I get to deal with so many half-abandonned/half-finished features.