MIVELD -

First, before thinking about linear time trend or exponential smoothing or
some other functional form, "look at the data." That is, plot monthly usage
versus time. Since you have three years of monthly data, you should be able
to see if there is a repetitive seasonal pattern. If so, you should include
the expected seasonal variation in your forecasts. Chapter 20, Time Series
Seasonality, of my book, Data Analysis Using Microsoft Excel, describes
three methods: regression with indicator variables, autoregression, and
classical time series decomposition.

- Mike
www.mikemiddleton.com

"MIVELD" wrote in
message ...

Hi I have recently been given the task by my boss to forecast one of our
key customers monthly usage until monthly until end of 2006,

I am starting with this customers indidual branches monthly usage for
the past 3 years, Some branches will have opened and closed during this
time, And i am looking to be able to predict monthly usage for the next
18 months

Has anyone got any idea's on the best way to forcast within excel, I am
currently using trend (fomulae given to me on this sight) but people are
saying this is not the best way

Exponentionally has been mentioned by a few of my collegues but i have
no idea of the difference between that and linear.

Thanks

Al


--
MIVELD
------------------------------------------------------------------------
MIVELD's Profile:
http://www.excelforum.com/member.php...o&userid=24562
View this thread: http://www.excelforum.com/showthread...hreadid=381536