Activity 7 - Predicting Order Problems

One of the biggest problems with predicting order problems is just that - the predicting. As its only a prediction, it is not garenteed to be right, and as such any specific change in the shopping habits of regular customers (e.g. decide to buy next years christmas presents in January so as to take advantage of any discounts availible) will have a significant effect on the model it vast numbers of customers do it. For one, if they were using the prediction, they would be lacking in January and need more emergency orders whilst being overstocked in Decemeber.

Due to these possible changes, the company will be losing out of profit by ordering emergency orders, along with being out of stock. This is the problem with the "order to supply" attitude, as changes in the consumers buying habits on vast scales can have incredibly damaging effects in theory.

Data Sources

Sensors can provide I decent accurate stream of data, allowing for predictions. With F1 cars for instance, the pit crew get a constant stream of data updated every second, allowing for them to plan and change their plan according to what data they are recieving.

The internet has a huge potential as a data source, with ease of accessing any data simply by using search engines however, it is always possible that data can be false. Printed word also does not have the ability to update its data easily, so it is often unsure as to how accurate the information actually is.

Surveys are useful, however due to sample size they are often only predictions of the whole populations response and can sometimes be wrong.