Sure! We operate in the toys and games space, but regardless of the space I'd follow the same process:
1) Locate a mid-popularity product with a ranking of < 4.3 stars and what appears to be profitable margins
2) Look over the neutral and negative reviews looking for improvement suggestions and complaints customers have
3) Figure out an improvement or a way to fix the issues
People will tell you what they want in the reviews. The hard part is figuring out a way to fix it for a price they are willing to pay. Price wins on Amazon every time, unfortunately, so you have to be prepared to sell for less or equal to the competition while at the same time having a superior product. We sell a product for around $20 when the competitor is priced around $18. Our product is 4.9 out of 5 stars with ~400 reviews while the competitor has significantly more reviews but only around 4.5 stars. It's obvious to customers that they're paying a slight premium for a better product when they purchase from us, so we're able to compete. That being said, there's one direct competitor that sells a 3.8-star product for $12 that outsells us 4:1, showing once again that price wins on Amazon's platform.
If you have a superior product, the rest will take care of itself. The marketplace may change and the market may change, but if you have a great product and you offer amazing customer service you'll be fine.
Of course, I'm glossing over tons of details here, but that's the gist of it. You're welcome to ask questions here or follow up with a PM. I'll be glad to help in any way I can.
I think the case of the 4:1 product at about half the cost but wall less positive reviews also calls in to question how effective Amazon's presentation algorithms are.
A logical solution would be to have their ranked search cherry pick the highest (number and score) reviewed results for focus within a given price category.
However I normally hate their ranked search since it's results normally don't seem relevant to my needs. Thus I tend to manually do that by sorting by price and then eye-ball filtering for quality/look in picture (yeah, I used that as a quick filter for poor quality).
It would be nice if I could instead directly assign weights to different type of metrics and have Amazon present the list according to my score based on perceived importance. They could use the data from that and actual buying outcomes to tune their default filter (that one might better balance high scores on multiple metrics being sorted to the top of the list).
