What should you upload and not upload?

If you have 5 models finished, why not just upload all of them? Yes, it's true that certain categories will tend to sell better than others, but it also depends on the quality of the models and how often you can add new content to your store.

However, if you only have 1 single model in your portfolio, your chances of making sales are minimal even if it's a good model. Think about the fact that there are probably thousands of good models in every category already. Start out by modeling what you know best and what you like to work with. As your portfolio grows and you analyze the analytics more closely, maybe then you can decide on a different category to focus your efforts or maybe stick with what you enjoy.. Just don't expect sales to happen quickly with only a few products. A single model could take months before someone discovers it and decides to purchase it.

EDIT: I just checked your portfolio. It seems you already have 5 models uploaded? For some reason, I only saw 1 model the first time. Sorry if I misunderstood.

I can say that your models has almost no value in the marketplace. They are very simple, even most inexperienced modeller will easily make them in a few minutes. Ask yourself why anyone should bother with your models if it would take him faster to make his own, than he could download it from here? If you want to have any chance to sell, you need to start making more complex models. Something that would save time and effort to customer and save his money.

@unscriptablee-rbxx single complex, good quality, adequately priced model will earn you much more than thousand boxes, spheres and cylinders.

Do you want to guess which of the two sellers makes more money by the end of a month?
https://www.cgtrader.com/kasiopy
https://www.cgtrader.com/creative3ddesigns

IMHO CGT's analytics tools have more "cons" than "pros". Just search the forum for questions and disputes about analytics. And read agreement... I made wrong step when I filed for CGT analytics. Now I can't to manage my models as it was possible in "free of analytics" life. And declining from CGT's analytics program looks hilariously complicated an uncomfortable.

Just do good models and publish those models as regularly as you can. And that's all. Also self-promotion on social networks and thematic forums can be useful. But I'm sure at this moment promotion on social platforms also more wastes the designer's time than gives good reflection on sales. (IMHO again)

I must say, I do find the analytics a useful tool when deciding what NOT to make next. That's where its importance lies.

I find the lack of movement of keywords in the analytics disturbing. They barely move up or down over time.