We've always had some trouble with logo design. Sometimes they are a big hit. Other times we struggle to come up with something the client likes.
I finally made a breakthrough on this, and,of course, it turned out to be a Homer Simpson moment ("Doh!").
We've long had a system on website designs to have the client send us links to some websites they like, along with what they like about them.
That's because when you are talking about visual preferences, they are very difficult to communicate completely in words. MUCH easier to say, see this? I like the contrasting colors. Or that swooshy shape.
That is even more true with logos.
All successful websites are going to have things in common about their looks. There are technical limitations (they are always going to be on a rectangular grid, unless done all in Flash). Also you need to work within usual expectations (menus are usually horizontal, near the top) or you lose people with confusion.
Logos are more different than websites. So much so that there are five complete classes of logos plus combinations.
So (as I wrote) logo design should start with a decision about what TYPE of logo it is going to be.
But then it should go into "what are some logos you like, and what do you like about them."
THEN logo design can be a smooth process.
Many design companies deal with this by offering you a certain number of complete designs to choose from. But what if none of them work for you? Eat the fee and hire someone else? Complain until they try again? Pick the least worst choice?
Much better to have a process that really nails it.