Expectation: I'll be able to use Google and the Mendeley web importer (bookmarklet) for everything.
Reality: When you use web importing, you often have very basic, or even incomplete, citation information. This is because webpages are not designed to give citation metadata.
Fix it: You can import website information as is, and add additional information manually. Do this as soon as you import, or you may forget. With peer-reviewed articles that you find on Google Scholar, you can try a couple of options.
One, make sure your Google Scholar settings allow for BibTex export, and use that feature to get the information into Mendeley, rather than the bookmarklet.
Two, instead of Google Scholar, search for your article from a database on the UGA Libraries website. Many databases are provided through GALILEO, and contain records dense with metadata. You can search for subject specific databases (like ERIC or PsycINFO), or use the very broad Multisearch. Chat with a librarian if you need a pointer or two.
Expectation: I can drag and drop PDFs into Mendeley and get all the citation information that I need.
Reality: This is a great feature of Mendeley, but it works best with the very recent PDFs that come from databases. Databases add lots of metadata to PDFs that citation managers find and use. If you have older PDFs, such as from JSTOR, or items that colleagues have scanned, the PDFs just don't have citation information to find.
Fix it: Send database records for your item to Mendeley. Merge duplicates, or attach the PDFs after the export from the database.