An inverse correlation has been found to exist between the magnetic flux tube expansion factor (FTE) at the source surface and the solar wind speed (SWS) observed at Earth, which has been made use of in the prediction of solar wind speed near the Earth with reasonable accuracy. However, the correlation between FTE and the solar wind speed at Earth is not always consistently high, and the discrepancies between the observed and predicted SWS are quite significant. There are several factors causing this discrepancy. The present work is an investigation of such factors and an estimation of error caused by them. We found that the number of multipole components included in the spherical harmonic expansion in the potential field source surface model, Nmax, has tremendous influence on the location of photospheric foot points of coronal features as well as the computation of FTE. Another important factor is the transit time of solar wind used in the inverse mapping. Approximate values of this parameter that are currently being used can lead to significant errors in the predicted solar wind speed. We also present a detailed discussion on the importance of considering interaction between slow and fast solar wind streams in the inverse mapping technique.