how to forcast sales data(other techniques besides moving averages?)

I'm new to the idea of forcasting, so want to get some inputs as I'm just experimenting with the data and going through a beginners book.

I have sales data from Jan 1 through March 26 (excluding Sundays since sales are 0).
I did a 3 day moving average, 7 day moving average, and 10 day moving average of my sales. To see which Moving average is the most accurate, I did a sum-of-squared difference between the actuals and the averages (I,e =SQRT(SUMXMY2($C$5:$C$73,$D$5:$D$73)/69)). The results were 166, 20, and 22 for 3 day, 7 day and 10 day. So that tells me that the 7 day average is the most accurate among the 3 (since it has the lowest difference)

However, looking at the individual average and comparing to the actuals, it's still off by quite a bit. How would I proceed from there? What other tools or methods are available that I can use to try to get closer?


Fortran must die
There are many alternatives, but three months of data is not really enough to generate results accurately. You can not adjust for seasonality for one thing. You really need several years worth of data.

Exponential smoothing (Holt Winston is one common form) and ARIMA are commonly used for forecast like this. There are also econometric models I am sure, but I do not work with those.