Help file for the =FORECAST() function.
https://support.office.com/en-us/art...4-7ad38bbeda99
As far as mechanics and syntax, yes, this is how the FORECAST() function is intended to be used. However, it appears to me that you need to review and understand the statistics behind the problem and the FORECAST() function to decide if this is really the analysis you want.
As explained in the help file, the FORECAST() function performs a simply "linear" regression to the standard y=mx+b equation. Then uses that result to compute y for the desired x value. The regression is naturally going to be sensitive to what you put in for the "known_x" values, as you have observed. Having the week number "cycle back" to week 1 probably does not make sense in this regression. You need to decide where in your "calendar" is week 0 and number the weeks forward and backward from there to get a meaningful "time" measure. I also note that your y values appear to vary on more of a "cyclical" pattern (high then trending down to a low then trending up to a high and so on). A regression that can account for this kind of cyclical behavior may be more appropriate. You will need to think through these kind of issues to decide how best to proceed.
Bookmarks