Why Do Forecasts Fail?

Why do forecasts fail? How do you recommend improving the results? What tools would you use? These are very important questions that you should ask yourself when making, monitoring, and updating a forecast. The answers to these questions will help you make a more accurate forecast or help you update or fix a forecast that may already be in place. Forecasts in their own nature are expected to have some type of error but with the correct techniques it can be measured and monitored.

Hire a custom writer who has experience.
It's time for you to submit amazing papers!


order now

Some factors that contribute to forecast error include: inappropriate forecasting method, lack of participation and accountability, too difficult to understand, lack of compatibility between system and organization, inaccurate data, data inappropriate for forecasting, and lack of monitoring. For example, monitoring the forecast is a very important task in order to track actual demand against projected demand and yet there are still companies that do not check the accuracy of the forecast.

It baffles me to think that a company would invest the time and money to create a forecast but not use it for what it is intended to do. This could create a big problem especially since there is no way to establish targets for improvement. Improving the results of a forecast takes time and knowledge of all factors involved. Developing a forecast process would be the first initial step that I would take before making any forecast. Taking this step will help to assure that I have all the required information and all the right personnel involved in order to develop a good forecast.

It is also important to have a good forecasting software program that will work with the order entry and the master scheduling programs as well. If I am working with a forecast that is already in place but needs to be adjusted to improve the results. I would focus on reducing forecast error, it will then increase forecast accuracy. Although both factors accentually mean the same thing if you focus on reducing the forecast error it makes it easier to fight bias, which is the worst type of forecast error. Bias in a forecast either high or low will typically cause lots of problems with supply.

A constant high bias in a forecast can result in layoffs, excess inventory and capacity, and more. On the other hand if the forecast is routinely low it would result in late shipments, unhappy customers, unplanned overtime, and on and on. You must measure bias, track it, report on it, as well as do anything in your power to minimize it. You have to be able to locate the root cause; there are many categories that it could fall into. For example, forecasting low in order to beat the plan, forecasting high to match the original plan despite the changing circumstances.

Tracking signals are a great tool used to relate variability (MAD) and bias so you can identify what type of error you are encountering. Another monitoring technique is the use of demand filters, which is a quantity limit setting. Since accuracy or error can be viewed at different levels I would use some or all of the following tools: • Period Forecast Error, this determines how much a forecast deviates from the actual demand for a given period. The formula is Forecast Error = Actual Demand – Forecast Demand. Absolute Percentage of Error, this can be found using the following formula APE = ¦A-F¦/ A x 100% APE determines how much a forecast deviates from the actual demand for a given period in percentage terms. • Mean Absolute Deviation, this is the average amount by which the forecast is in error no matter what direction. It can be calculated using MAD = ? ¦A-F¦/ n or MAD = Sum of absolute errors/ Number of periods. • Mean Absolute Percentage of Error enables us to know how much the average percentage forecast values vary from actual values.

It can be found using the following formula, MAPE = ? {¦A-F¦/A} % /n. •Standard Deviation is pretty much the same technique as MAD so it measures the amount of error and does not take into account the direction of the error. The correct formula is Standard deviation = v? (Actual – Forecast) ? /n-1 There are many tasks that can be looked into in order to fix a forecast but I would try to focus on making it right from the beginning. There are a lot of steps and information needed to develop a good forecast.

Even when you have all the information if you don’t have the participation or the support from the right personnel it will be very difficult to achieve. Once a forecast is made and in place it is just as important to monitor as it is to create. The forecast must be watch closely and corrective action must be taken outside of random error. If you do not follow the proper procedures in monitoring the forecast you might as well not waste the time in developing one. It will be of no benefit to have a forecast and not know if actual demand is in line with it or not. Sherry Kennedy

x

Hi!
I'm Heather

Would you like to get such a paper? How about receiving a customized one?

Check it out