| Course | BUS 352 Business Statistics |
|---|---|
| Topic | Topic 8 |
| Paper type | Management findings report |
| Length | About 790 words, 1 pages |
| Format | Memo with APA 7 references |
| School | Grand Canyon University (GCU) |
| Program | BS in Business Management |
| Updated | September 2026 |
Free sample paper for BUS 352 Topic 8
Memorandum
To: Vice President, Customer Operations, Keystone Utilities
From: [Author Name], operations analytics
Date: [Date]
Re: Results of the callback system pilot
Recommendation
Buy the callback system for all 24 contact centers, on one condition: that the vendor agrees to a performance clause tying the second-year license fee to the abandonment rate staying at or below the pilot result during the winter peak, which the pilot did not cover.
The Headline Result
In the six centers randomly given the callback system, the share of callers who hung up before reaching an agent fell by about three percentage points compared with the six centers that were not given it. The plausible range for the true effect runs from about one point to about five. For a center handling 18,000 calls a month, three points means roughly 520 fewer customers a month giving up on the call. Customer satisfaction scores did not change detectably: the average score in both groups moved by less than a tenth of a point on the five-point scale, well within ordinary month-to-month variation.
How the Pilot Was Run
Twelve centers of similar size and call volume were paired, and a coin toss chose one center in each pair to receive the system. Random assignment means the two groups differed, on average, only in whether they had the system, so the difference in their results can be attributed to it (Kohavi et al., 2020). Abandonment was tracked in all twelve centers over the quarter before installation and the quarter after it. In the treated centers it fell from 11.0 to 7.9 percent; in the comparison centers it moved from 11.4 to 11.2 percent. The difference between those two changes, 2.9 points, is the estimated effect, and at 95 percent confidence the true effect lies somewhere between roughly 1 and 5 points (Levine et al., 2020).
Table 1
Abandonment Rate Before and After
| Before | After | Change | |
|---|---|---|---|
| Centers with the system | 11.0% | 7.9% | -3.1 points |
| Centers without | 11.4% | 11.2% | -0.2 points |
| Estimated effect | -2.9 points, range about -1 to -5 |
What the Satisfaction Result Means
It would be easy to leave the satisfaction result out, since it is not good news. It is reported because it matters: the system reduces hang-ups without, as far as the pilot can tell, making customers who do get through any happier. That is not surprising, since callers who wait for a callback still have the same problem to solve. Not finding a change is not the same as proving there is none; a small effect on satisfaction could exist that six centers cannot detect (Wasserstein & Lazar, 2016). The case for the system rests on fewer abandoned calls, not on satisfaction.
Why Randomize Instead of Choosing the Centers
Operations managers originally proposed installing the system in the six centers with the worst abandonment, since they had the most to gain. That design would have made the results impossible to read: centers with unusually bad months tend to improve on their own the following quarter, so their improvement would have looked like the system's effect even if it had none. Assigning the system by coin toss within matched pairs avoided that problem, and it is why the three-point estimate can be trusted as the system's effect rather than a mix of the system and ordinary fluctuation.
What the Pilot Cannot Show
Four things are outside what the pilot measured. It ran from April to June, when call volume is moderate; performance during the winter peak, when outages and billing questions double volume, is untested, which is why the recommendation carries a condition. It did not measure effects on agents' productivity, such as whether scheduled callbacks crowd out live calls at busy times. It lasted three months, too short to show whether customers who use callbacks are more likely to stay with the company. And it did not test the costs of integration at the older centers, which use a different phone system.
Cost and Payback
The system costs about $14,000 per center to install and $6,000 a year per center to license. Each abandoned call that is later recovered costs the company an estimated $4.50 in repeat calls and field visits. At about 520 fewer abandoned calls a month per center, the savings of about $28,000 a year per center cover the costs within the first year. The estimate is rough; even at the low end of the effect range, one point, the system would pay for itself in about three years, still within its expected life.
Next Steps
If you approve, operations will negotiate the performance clause, schedule installation center by center beginning with the largest, and continue measuring abandonment and satisfaction monthly, including through the winter peak. A report on the peak-season results will follow in March.
References
Kohavi, R., Tang, D., & Xu, Y. (2020). Trustworthy online controlled experiments: A practical guide to A/B testing. Cambridge University Press.
Levine, D. M., Szabat, K. A., & Stephan, D. F. (2020). Business statistics: A first course (8th ed.). Pearson.
Wasserstein, R. L., & Lazar, N. A. (2016). The ASA statement on p-values: Context, process, and purpose. The American Statistician, 70(2), 129-133. https://doi.org/10.1080/00031305.2016.1154108
BUS 352 Topic 8 instructions, in plain terms
The Topic 8 instructions in BUS-352 ask students to report statistical findings to management. Weak submissions present the analysis in the order it was done. The assignment rewards reports that lead with the recommendation, give the headline result in the reader's own units with its plausible range, and explain briefly how the evidence was produced. Students should report results that did not come out as hoped, state what the study cannot show, translate the effect into costs and savings, and propose next steps that address the gaps, all written for a reader with little time.
How the BUS-352 Topic 8 example is put together
The finished report fits the attention of someone deciding between meetings. Its first sentence is the recommendation: buy the callback system for all centers, subject to one condition. The second paragraph gives the headline result in the reader's own units, from illustrative pilot data: in the six centers randomly given the system, the share of callers who hung up before reaching an agent fell by about three percentage points compared with the six that were not, and the plausible range runs from one point to five. Customer satisfaction scores did not change detectably, and the report says so rather than leaving it out. A short list names what the pilot cannot show, including performance in the busy season. The cost comparison uses the low end of the range. A technical note carries the tests for anyone who wants them.
BUS 352 Topic 8 rubric: what Excellent looks like
Faculty grading the findings report reward a decision a busy reader can make from the first paragraph. Excellent reports express the effect in hang-ups avoided and give its plausible range. Credit goes to explaining why random assignment makes the result trustworthy. The rubric values reporting the unchanged satisfaction score rather than leaving it out. The highest level lists what the pilot cannot show and ties the recommendation's condition to the biggest gap. Reports that bury the recommendation after the methods usually lose marks with graders.
BUS 352 Topic 8 help: mistakes that cost marks
A findings report written for the analyst instead of the decision-maker is the defining failure of this topic. Opening with the method, the hypotheses and a table of test statistics forces the vice president to dig for the answer, which a busy reader may never do. Jargon costs marks even when accurate: rejecting the null means nothing to someone who needs to know how many calls the system saves. The opposite loss is a report so simplified that the uncertainty disappears, stating that the system reduces abandoned calls by three points as though the range did not exist. Omitting the satisfaction result, because it did not help the case, is a quiet form of misreporting that careful graders catch. Recommendations with no condition, no cost comparison and no follow-up measure leave the reader nothing to hold the decision against.
Get a BUS-352 Topic 8 example written to your instructions
Send the BUS-352 Topic 8 instructions, the rubric your classroom posts and the analysis or output your report has to communicate. We write a custom example to those criteria, with the recommendation first, the result stated in the reader's units with its range and the technical detail kept out of the way, delivered in 24 to 48 hours. The first one is free.
More BUS 352 papers and BS in Business Management samples
- BUS 352 Topic 1: Descriptive Statistics Memo
- BUS 352 Topic 2: Distribution Shape Analysis
- BUS 352 Topic 3: Sampling Variability Exercise
- BUS 352 Topic 4: Interval Estimate Memo
- BUS 352 Topic 5: P Value Dq Response
- BUS 352 Topic 6: Two Sample Test Report
- BUS 352 Topic 7: Simple Regression Interpretation
- BUS 340 Topic 8: Position Defense Paper
- BUS 332 Topic 8: Targeted Retention Plan
- BUS 317 Topic 8: Funding Request Memo
BUS 352 Topic 8 questions, answered
How do I describe uncertainty without statistical terms?
State the result as a best estimate with a range and say what the range means for the decision. For example: abandoned calls fell by about three points, and the pilot is consistent with a true improvement anywhere from about one point to five. Then add the practical reading, that even the smallest likely improvement covers the license cost. That keeps the uncertainty visible without asking the reader to know what an interval is.
Where do the p-values and test results go?
In a technical note at the end, or wherever your rubric directs, so that anyone checking the analysis can find them and nobody deciding has to read them. The main text can mention that the difference was tested and held up, in those words. Some sections require the statistical results in the body as well, and where yours does, place them after the plain-language finding rather than before it.
Should the report mention results that did not support the recommendation?
Yes. A decision-maker who later learns that satisfaction did not improve, and that the report left it out, will discount everything else in it. Reporting the unhelpful result, and explaining why the recommendation stands anyway, is what makes the favorable result credible. It also protects the analyst, since the decision was made with the full picture rather than a selected one.
Where can I find a free BUS 352 Topic 8 sample paper?
The report is here in full: the recommendation, the headline result, how the pilot was run, why it was randomized, what the satisfaction result means, what the pilot cannot show, cost and payback and next steps. Share the pilot data and the decision it informs for a free worked report.
How large was the effect in BUS 352 Topic 8?
About three percentage points fewer callers hanging up, with a plausible range of one to five, or roughly 520 fewer abandoned calls a month per center.