Wow, thank you very much for your thorough analysis and feedback. It makes me really excited and motivated to know, that someone put a lot of thought into it. Yeah! Check what is character saying when he does fiverr, it's meant for joke :)
I agree it would be much better to capitalize more on item types. I'll try to explain how I thought the basis game loop may work: So at low levels, you have to think if you lose more money and buy expensive stuff, I will level faster because one part of experience equation, but you will probably have to invest some points to Fiverr gigs to compensate our money lost. On the other way, you can start selling cheap items and lost none or minimal amount of money, so you can invest points into attributes.
The reason why attributes sometimes seemed to not work is, that bargain chances are composed of two parts. The first one is how well you fulfill customer requests. The second one is computed from the difference between your attributes and customer attributes. My game doesn't communicate that in any way, so it would be much better to show customer attributes as well as his preferences. I consider this a design flaw.
How to best capitalize on customer preferences - if you cannot know what scores have the items you are buing from suppliers? One thing is that usefull and style points mostly go from type of item and age and luxury highly corelate with price. Yet, it's not 100% true. So the other way around is to buy anough showcases that you can hold to big inventory and have stuff to offer in every category. Things that doesn't fit that can be then sold to customers with no preferences.