Statistical Chocalate making - the saga is only three parts old

by Erica Keogh

Click here for Part 2 of this article

Welcome!

We are now at last at the final stages of production. What more is needed?

Sampling: In order for our packaging to be as nearly perfect as possible it will again be necessary to sample and check the packaging quality. Statistical tools often used at this stage include the Pareto chart and Defect Concentration diagrams, both of which enable you to monitor how many and what types of defects are most common.

Surveys: Reliable market research should involve correct statistical design of surveys to gain information on customer satisfaction and on advertising methods and places. The sampling method should be probability based, and the sample size should be sufficiently large, for valid generalisations to be made. The questionnaire should be pre-tested so that you ensure that the question you think you are asking is in fact being asked. The data should be correctly coded and analysed, preferably by a statistician!

Econometrics and Forecasting: Once again the techniques of econometrics, times series and forecasting, are essential for correct predictions of sales' volumes and, thereby, suitable prices. Modelling will complement the forecasts by taking into account all possible variations in the market.

Operations Research Products have to be delivered at set times and within deadlines laid down by the purchaser. It is impossible to properly coordinate large scale operations without the tools of queuing theory, networking, inventory management, and other such techniques.

Our chocolate bars should now be on the counters ready for the demanding public to savour and enjoy. What have we forgotten in our statistical treatment of this product? See this spot in the next issue for the answer and the final exciting instalment of Statistics in Action in Industry in Zimbabwe!