this can sometimes mean you are amplifying more than 1 specific product; invariably s second primer diner species: this will show as a second melt curve in your dissociation cycle at the end of your standard curve PCR cycle
Sounds odd but is a common feature of RT-PCR. Ideally, your efficiency should be 100% but often amplifications are somewhere between 95-105%. Deviations from the 100 in either direction reflect the highly dynamic processes underpinning the method, and primer dimers -as mentioned by Laurence- or minor unspecific amplification can cause such values. After all this method provides a good estimate of copy number, but the true value may be slightly different because of the method's routine experimental error.