Data has been gathered on the time (minutes) to complete the most critical
step in an invoice processing operation. See the data file.
The requirement is to complete the step within 15 minutes.
(a) Is the procession time for the step stable?
(b) Is the procession time for the step capable? How often will the target
(c) What is a plausible range for individual step times?
(d) What is a plausible range for the average step time?
(e) Suppose going forward into the next reporting period, the average time
was 10 minutes. How should this result be interpreted?
(f) Suppose going forward into the next reporting period, the average time
was 8 minutes. How should this result be interpreted?
(g) Is it necessary that the times follow a normal distribution for the
results for parts (d), (e) and (f) to be valid? Explain.
Page limit: Maximum 4 pages including all diagrams.
Group Assignment (2015) 3
Question 2 30 marks
The team leader of an accounts payable team was concerned about delays in
paying vendors so she decided to baseline performance. The team randomly
selected 5 invoices each day; one each at 9:30am, 11am, 12:30pm, 2pm and
3:30pm and monitored time to pay vendors over six working weeks
comprising 30 days. The data are labelled as ‘Before’ in the data file.
The team leader wanted to monitor the variation within each day and
between the days.
The company has a policy of paying vendors within 30 days of invoice
receipt. The company was concerned how often it was keeping its promises.
Suggestions for process improvement were brainstormed with subject
Improvements, which were expected to have an immediate effect, were
implemented over six working weeks comprising 30 days and the data
labelled as ‘After’.
Describe the ‘Before’ performance using the appropriate control chart(s).
Identify and remove all special causes. Provide a summary of the final
control chart(s) at the end of your analysis.
Assess process capability with respect to the 30 day policy.
Describe the ‘After’ performance using the appropriate control chart(s).
What can you conclude? You must restrict your analysis to the use of
control chart(s). (There is no requirement to compare ‘After’ with ‘Before’.
Simply assess ‘After’ on a stand-alone basis.)
You are particularly interested in whether the average time has dropped
between the ‘Before’ and ‘After’ processes. Using only graphical and
numerical summaries found in Unit 1 report on the extent to which the
average time to pay vendors has improved with regard to paying vendors at
or within 30 days of invoice receipt.
Page limit: Maximum 9 pages including all diagrams.
Data Analysis & Statistical Modeling for Business (2015) 4
Question 3 25 marks
The Chief Information Officer (CIO) of a large multinational firm has asked
for your assistance to quantify improvements in repeat call rates (RCR) in
its customer hotlines across the world. The firm recently rolled-out
expensive new information technology in each of its 50 sites to assist
hotline operators reduce RCR.
A call is resolved on the ‘first call’ when the operator can answer the
customer’s query during the first phone call and does not need to ring the
customer back at a later time with the solution. The customer does not need
to call back either because the operator has solved the customer’s problem
on the first call.
A ‘first call’ is effectively a ‘closed call right first time’. A ‘repeat call’ is a
‘call that generates one or more additional calls before it is closed’.
The new technology is aimed at reducing the causes of repeat call rates
(RCR) by providing operators with more in-depth information on issues
You are given data on RCR before and after the installation of the new
technology in each of the 50 hotlines. See the data file.
Repeat call rates (RCR) are expressed as the average number of repeat calls
per one hundred customer inquiries. Data was collected over one month
before and one month after the new technology was introduced and training
As part of a cost-benefit business the CIO wanted to know if RCR ‘after’
the installation of the new technology for each site could be predicted from
the ‘before’ situation. The CIO stipulated that any model used for prediction
should be as simple as possible yet validity of the predictions should not be
He was particularly interested in Site 49, which historically had been
amongst the worst performers. In Site 49, the CIO decided to change the site
manager as the new technology was being introduced.
He also wanted you to assess the extent of the improvement. He was hoping
to get an average improvement of more than 10% across all those sites
remaining in the predictive model.
Your submission should be in the form of a report to the CIO with the first
major section titled Exploratory Analysis where no model is built but the
data is investigated and the second major section titled Regression Analysis
where the appropriate model is built.
In your analysis, ‘RCR Before’ should be on the X axis and ‘RCR After’ on
the Y axis.
Group Assignment (2015) 5
You may need to make and justify decisions on what data is retained and
what data is removed.
The full Reporting Template is NOT required to summarise your model and
you are only expected to write out the equation of the regression model and
discuss its explanatory power.
You will also need to identify outliers and leverage points and use the
The third major section should be titled Extent of Improvement and test to
see if there is an average improvement of more than 10% across all those
sites remaining in the predictive model. Discuss any shortcomings of your
analysis so that the CIO is fully informed of your results.
The extent of improvement should be expressed as a Reduction % where:
ï€¨RCR Before RCR After ï€© ï€¨ ï€©
Reduction % 100