Kay and Michael Passe publish \"What\'s Happening? -- a biweekly newspaper to pu
ID: 3040206 • Letter: K
Question
Kay and Michael Passe publish "What's Happening? -- a biweekly newspaper to publicize local events. What's Happening? has few subscribers; it typically is sold at checkout stands. Much of the revenue comes from advertisers of garage sales and supermarket specials. In an effort to reduce costs associated with printing too many papers or delivering them to the wrong location, Michael implemented a computerized system to collect sales data. Sales-counter scanners accurately record sales data for each location. Since the system was implemented, total sales volume has steadily declined. Selling advertising space and maintaining shelf space at supermarkets are getting more difficult.
Reduced revenue makes controlling costs all the more important. For each issue, Michael carefully makes a forecast based on sales data collected at each location. Then he orders papers to be printed and distributed in quantities matching the forecast. Michael's forecast reflects a downward trend, which is actually present in the sales data. Now only a few papers are left over at only a few locations. Although the sales forecasts accurately predict the actual sales at most locations, What's Happening? is spiraling toward oblivion. Kay suspects that Michael is doing something wrong in preparing the forecast but can find no mathematical errors. Tell her what is happening.
1) What is the root cause of the problem Michael is facing?
2) What is the appropriate solution to the problem?
3) Once the problem is solved, what types of forecasting approaches might Michael use and (most importantly) how should he decide which one to ultimately employ?
Explanation / Answer
Answer (1) : The root cause of the problem is that of Sampling Bias, which means that the data which is being used to make forecasts is not sufficiently representative of the true population, but is systematically biased toward only a small part of the true population.
In this example, after introducing the computer system, it is known that the sales volume has steadily declined. Michael is using this data alone in order to make new forecasts. Unfortunately, the data collected after introduction of computer system is biased because it is representing only that part of the population which is affected by the introduction of computer system. So Michael will get forecasts which are going to be showing downward trend. And then he prints exactly those many number of papers as predicted by this downward trending data. This makes a continuously declining sales volume due to the vicious cycle created by Sampling Bias.
Answer (2) : Appropriate solution to the problem would be to use both (i) The data collected after introduction of computer system adn (ii) data collected prior to the introduction of computer system. If we use this collective data, then our forecasts will be much more reliable and will certainly not keep giving lesser and lesser sales volume forecasts, which in turn will help in growth of the "Whats happening?".
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.