In a business context it is common to use ‘month’ as a time unit and compare a month to its predecessor or to the same month of last year. So it is important to at least keep in mind that months can be very differently seasonally biased. Most  business dynamics exhibit a week based seasonality with specifically strong weekdays. March 2011 for example has 23 work days – 4 more than the following month (holidays considered for Bavaria). This can have a considerable effect – especially when coincidentally leveraged by the usual volatility.

How to deal with this issue depends on the inner workings of the respective numeric data and the way it is intended to be used eventually. Something like an average for the work days could be an idea. In my opinion month as a time unit is pretty much flawed to begin with as it suggests a not existant periodicity. But as long as one keeps this in mind wrong conclusions can be avoided.