If there was an unusually high peak of seasonal jobs just prior to the months in question, that would be an utterly meaningless statistic. If it's a moderate random drop coinciding with the end of a somewhat high seasonal number, it's still not relevant to long-term outlook. That's why monthly numbers are difficult to use in isolation.