Manging the Springfield Herald Essay
Sorry, but copying text is forbidden on this website!
SH 13.1 – What criticisms can you make concerning the method of forecasting that involved taking the new subscriptions data for the prior three months as the basis for future projections?
– 3 months is a valid for comparing quarterly growth retroactively, but more time is needed for a valid future forecast. – Are there seasonal quarters historically where sales are higher?
Also, the factors that have affected the sales in the past three months must also remain constant for the entire year or at least the average sale/hour should be consistent.
Here the average sale/ hour for the prev months were as follows: Month 1= 4 ; Month 2 = 3.89 ; Month 3 = 5.1 . Again, the total average of the three months is: 4.32 which has an error variation of -0.8 to 0.4. This range is comparatively very high, as, for a three month data we would expect the error to be close to zero.
SH 13.2 – What factors other than the number of telemarketing hours spent might be useful in predicting the number of new subscriptions? Explain
– How many trial subscriptions are currently in circulation/ what is a normal rate of trial subscriptions converted into regular descriptions.
– What is the percentage of the population within distribution that does not already take the Herald (i.e. potential target customers)
Types of subscription: Annual, quarterly, monthly, weekly
Number of calls made
Turnover rate of the regular subscriptions
Post subscription services
Number of premium members
SH 13.3 (A) – Analyze the data and develop a regression model to predict the number of new subscriptions for a month, based on the number of hours spent on telemarketing for new subscriptions.
We can start by making the following analysis:
Take the data as it is given
Rearrange the data according to the number of hours and then predicting it Group them according to quarterly hours and then looking for the possibilities
For (i), we get a durbin Watson statistic that is >2 and the analysis is refuted. For (ii), we are getting it <2 and we can continue the analysis and further estimate the numbers
SH 13.3 (B) – If you expect to spend 1,200 hours on telemarketing per month, estimate the number of new subscriptions for the month. Indicate the assumptions on which this prediction is based.
This assumption is valid. because we can predict within the relevant range of data When using a regression model for prediction. 1,200 hours is within the range of data.
SH 13.3 (C) – What would be the danger of predicting the number of new subscriptions for a month in which 2,000 hours were spent on telemarketing?
– 2000 hours are outside of the range of the data (1498 being the largest in the range). Therefore, the data does not have the necessary information to perform a valid regression with any degree of confidence, in order to make the forecast.
We have the predicted data. It says: for 2000 hours the prediction is around 8000. The problem here predicting is that the uppermost level of the data we have is for 1500 hours appr and the predicted is more than 500 hours that needs prediction. As we move further from the established data the error of prediction is higher. Therefore, we can be 95 % confident of the predictions for 1500 hours. But since it is for 2000 hours the confidence interval would increase. ( I am not sure what the interval is but we will need to calculate it…will need some more time figuring it out). Moreover, we have considered linear trend for prediction which may not be true.