First of all, this > happens only if the alternative hypothesis has a > (and when you're testing means..)

The p-value tell you how far out in the tails your test statistic lies.

So if the p-value is really small, then the data indicates that the null hypothesis may be wrong.