I have to apply Bayes' Theorum to solve this question:
In 2001 a cider company's profits rose 55% over the previous year ($6
million in profits in 2001 and $3.87 million in 2000). Now, the
company needs to decide if they should buy more kegs, but their market
forecast will not be prepared for 1 more month. Their marketing
manager gives a "best guess":
"a 60% chance sales will be 10% up on August last year, 30% chance
they'll increase by less than this amount, and a 10% there will be no
increase". To get a rough idea on the profit their financialmanager
assumes:
1. either 0, 4000, or 8000 will be ordered;
2. if sales increase more than 10%, all extra kegs purchased will be used;
3. if sales increase less than this amount only 4000 extra kegs would be needed;
Next month's forecast 's quality is assessed by the marketing manager
as "not bad". Advise the managers.

Im confused on how to solve this using Bayes' theorum though; any recommendations?

The question isn't clear.  I could make a guess but it seems also that
some information is missing.

I assume that the question is how to maximize profits given a forecast of:

60% chance that sales increase 10% or more:  assume 10%
30% chance that sales increase more than 0% but less than 10%:  assume 5%
10% chance of no increase:  assume 0%

Additional information required:

How much does each additional keg cost and how much profit there will
be for each additional keg if it is sold.  Also, does it take exactly
8000 additional kegs to accomodate a 10% increase in sales.  If so, it
follows that an additional 4000 kegs will accomodate a 5% increase in
sales.