It just really depends on what the project is. If there were some generic way to evaluate all $500 donations, then some centralized organization would be doing that already. You have to use your own personal, human judgment.
Let's look at the example of your local rationality community wanting to host a big event. It seems like there's no room easily available for that size but you could spend $500 for room rent for the event.
Given that you have a better understanding of your local community and the value the event would provide them someone who evaluates grants at LTFF it would be a very inefficient process if such a project would be funded by you giving money to LTFF and then LTFF reading a grant review for the project and deciding based on the grant review without knowledge of the local community that the project is worth funding.
How would you go about using your judgment in that case?
In my opinion that's a small enough amount of money that if, for example, you already have a strong sense of something like EA principles and want to give in an EA manner, then you're probably well served by following your intuition that's been informed by EA values. I say this because I think most of the value of small donations lies in taking risks on high variance projects that are less likely to receive funding from major sources, and so most of the value you can extract with your donation lies in the hard-to-quantify things you know and believe that lead you to think a particular project or organization is worth donating to.
I think of this as something like creating an ecosystem where people maximize variance by maximizing their charitable "alpha" make it more likely that good things get funded that are outside the risk tolerance of large funding sources and on net results in more good than a more conservative approach (here my model assumes the main cost of giving to projects and orgs that don't deliver is opportunity cost and that they end up being net neutral rather than negative otherwise).
At habryka's recent office hours he made the point that it's really hard for a big grand-giving organization like the Long-Term Future Fund (LTFF) to fund certain projects. While good rationalist culture for example is valuable, funding culture by committee in a top-down way is problematic. The people with local knowledge of a cultural project are in a better position to judge whether or not a given rationalist culture project is valuable enough that it should be funded then a grant committee like that of the LTFF.
Given that LTFF and OpenPhil have access to a lot of financial captial these days for very rich individuals, the scare resources these days isn't primarily money but local knowledge about which projects are worthwhile. This suggests that individuals who have a small donation budget and access to a lot of local knowledge of individual projects are more effective when they make use of their scarce resources of local knowledge instead of donating to money to big grant making organizations.
While giving money to big grant makers is straightforward, giving money to small projects is harder. How should an individual who has a yearly donation budget of <$10,000 go about evaluating local projects for donating an amount like $500 to them?