A Simple Look at Business Forecasting

If you have ever thought about making a business plan but you have little financial or statistical background, chances are you've gotten bogged down in forecasting. Forecasts are essential to every business plan because they help to determine what will happen in the future and what you will need to do to ensure that it happens correctly. Without forecasting your business will just be guessing from day to day and will most likely always be a step behind, something most business can't afford. In this article I will give you a simple look at forecasting, how its done, and what it means.

Forecasts are based on many variables and acan be used to predict many things like output and capacity. No matter what you're forecasting the basic concepts remain the same. In general a forecast is made using a time series. A time series is a set of observations of data (wether its your monthly sales or your daily output rate) measuring the same thing for consecutive time periods of the same length. Time series give us the data on past performance of our variables and how that performance has changed over time. It can allow us to spot trends or simply project what will happen in the future.

Time series patterns can show us five different things about the past when graphed as a line. If the line is horizontal that shows us a fairly steady flow of demand, we can use this to forecast by simply taking an average of our past observations. A diagonal line shows us a trend. A trend can be up or down and should skew our forecast in that direction unless we plan corrective action (in the case of an unwanted downward trend). Seasonal trends will show up as a line which goes up and down once or more each year. This shows a our business is seasonal and future forecast should take the month into account heavily. Finally it can be cyclical. Cyclical and seasonal are nearly the same but the ups and downs in a cyclical time series will take place over longer periods of time. Finding a pattern here is key or forecasting can be difficult.

You must make a few key decisions before you forecast. First decide on your units of measure and keep them constant. Decide next on the time series your checking, make sure that is relevant (so that each forecast gives you time to react, you wouldn't want to just forecast for tomorrow since it gives you no time to compensate for things if you need to). The time series you choos plays heavily into your "time horizon". This is essentially the time it would take you to react to the new data.

With all of this in mind you should be able to take on some forecasting. Now when you look at excel models asking for dat you will know what they are asking for and what the output means. You may not be able to create a forecast on your own, but you can hire someone for that later when your business is up and running!