Oh, I think I forgot to answer the basic question. The more powerful log analysis software packages should be able to give more granularity. Although it won't address simple questions without a knowledgable person doing the software configuration.
What constitutes a visit? Here's an example of how difficult it to determine. I sit behind multiple T3s; I download a PDF file. Oops, time to go to a staff meeting. Come back 30 minutes later and continue surfing - I'll read the PDF on the train.
How do you process daily or hourly statistics? If I surf to your site at 11:55 P.M. PST and don't download another page until 12:02 A.M. the next day, is that one hits or two?
These basic sort of statistical decisions may radically alter the end results. Edward Tufte's book Visual Explanations describes some of the pitfalls in a much more elegant manner than I've described above.
Again, the onus is on you to provide the statistical reports that your client desires. With zero knowledge of the data that you are to report on, it is a very tall order for others to propose a workable, satisfying solution that your client would be happy to pay for.
