I am trying to create a forecast using a monthly timeseries data set of marketing expenses for a fictional company. The data looks something like this:

Marketing Expenses Data Set

Using linear regression to forecast future sales, I get the following result:

Linear Regression

My problem lies with the seasonality of the marketing expenses (higher in the summer months for instance). I would ideally like to calculate to forecasted values of future months including seasonality. I read somewhere about ARIMA forecasting, but am really searching for some best ideas on how to accomplish the task.

To be clear, I do not JUST want a chart and trendline, but the data to support it too.

Any help would be much appreciated!


(Seasonal) ARIMA would be one possibility, but fitting it with plain Excel without specific plugins would be a major pain.

An alternative would be Exponential Smoothing, specifically seasonal Exponential Smoothing, also known as the Holt-Winters method. The formulas can easily be coded in Excel.

I warmly recommend Forecasting: Principles and Practice, 2nd edition, by Athanasopoulos and Hyndman. And I also recommend that you consider looking at R. FPP2 is a good introduction to time series forecasting using R.

  • $\begingroup$ Thank you! I'll take a look at it. I have R knowledge, but unfortunately my work only allows Excel. Trying to convince them that it's free and not dangerous... $\endgroup$ – Dan Dec 19 '17 at 11:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.