| Course | BU 321 Business Statistics |
|---|---|
| Unit | Unit 8 |
| Paper type | Statistical decision memo |
| Length | About 840 words, 5 pages |
| Format | Business memo with analysis, APA 7 references |
| School | Herzing University |
| Program | BS in Business Management |
| Updated | September 2026 |
Free sample paper for BU 321 Unit 8
Memorandum
To: Owners, Brickoven Express
From: Operations Analyst
Date: April 29, 2026
Re: Recommendation: a separate delivery promise for orders 6 miles or more from the store
Recommendation
Keep the 35-minute delivery promise for orders under 6 miles, and quote a 50-minute window for orders 6 miles or more from the store, starting June 1.
The reason is that the 35-minute promise is kept for almost every nearby order but for fewer than half of the distant ones. In our October sample, 95 percent of orders under 6 miles arrived within 35 minutes, compared with 48 percent of orders 6 miles or more. Distant orders are about one in ten of all orders, so the change affects few customers while removing most of our broken promises. The analysis behind this recommendation appears in the exhibits below.
Exhibit 1: On-Time Performance by Distance
Source: random sample of 240 delivery orders, 40 from each store, October 6 to 19, 2025. On time means delivered within 35 minutes of the order.
| Distance from store | Orders | Delivered within 35 minutes | Share on time | Mean delivery time |
|---|---|---|---|---|
| Under 2 miles | 75 | 75 | 100% | 18.4 min |
| 2 to under 4 miles | 92 | 90 | 98% | 24.1 min |
| 4 to under 6 miles | 50 | 41 | 82% | 29.4 min |
| 6 miles or more | 23 | 11 | 48% | 36.4 min |
| All orders | 240 | 217 | 90% | 24.6 min |
Exhibit 2: How Delivery Time Grows With Distance
A regression of delivery time on distance, fitted to the same 240 orders, gives: predicted minutes = 14.8 + 3.1 times miles. Each additional mile adds about 3.1 minutes on average (95 percent confidence interval, 2.7 to 3.5 minutes). Distance explains about half the variation in delivery times (R-squared 0.50), and dinner orders run about 2.4 minutes longer than the line predicts. The model predicts 33.6 minutes for a 6-mile order and 36.8 minutes for a 7-mile order before any dinner-hour delay, which is why the 35-minute promise fails at the edge of our zones. The typical miss around the line is about 6 minutes, so a 50-minute window leaves room for a slow dinner-hour trip. In the sample, all 23 orders of 6 miles or more arrived within 50 minutes.
Exhibit 3: The Chain as a Whole Is Performing Well
The recommendation is not a response to poor overall performance. Across all orders the average delivery took 24.6 minutes, and at 95 percent confidence the chain's true average falls somewhere from 23.6 to 25.7 minutes. A one-sample test against the franchise standard of a 26-minute average gave strong evidence that the chain meets it (t = minus 2.60, p = 0.005). The on-time share for all orders is 90.4 percent, with a 95 percent confidence interval of 86.7 to 94.1 percent. A comparison of the Eastgate and Northside stores found no statistically significant difference in delivery times (p = 0.20), so the problem is distance, not a particular store; distant orders came from every store, from two at Eastgate to five each at Northside and Midtown.
Why a Separate Promise Rather Than a Smaller Zone
Two alternatives were considered. Shrinking every store's delivery zone to 6 miles would end the late distant orders but also end those sales. Distant orders averaged $29.58, almost the same as the chain's overall average of $30.50, so they are ordinary business, not small tickets. Adding drivers would shorten waits but not the drive itself; Exhibit 2 shows that distance alone adds about 19 minutes to a 6-mile trip compared with a trip next door. A separate promise keeps the sales and makes the promise honest. Customers usually judge a delivery against the time they were quoted, so a realistic window that is kept may satisfy more than a short one that is broken, though this memo does not test that directly.
Limits
Four limits apply. First, every store contributed the same 40 orders regardless of volume, so chainwide figures weight every store equally; if some stores handle far more distant orders, the true share of affected orders may differ from one in ten. Second, the sample covers two weeks in October, and winter weather or summer volume could change delivery times. Third, only 23 distant orders were in the sample, so the 48 percent on-time figure is uncertain; a 95 percent interval around it would be wide, roughly 27 to 68 percent, though every value in that range is far below the 95 percent achieved nearby. Fourth, the regression describes association, not a tested effect of distance alone, and the memo does not measure how customers will react to a longer quoted window.
What would change the recommendation: if a larger sample showed distant orders arriving on time at a rate near 80 percent, a single promise might be kept with better dispatching instead. The cost of finding out is modest; tracking every order by distance band from June through August would give several hundred distant orders and settle the question. This memo, the chain and its data are teaching composites, and the methods follow standard practice (Anderson et al., 2020; Wasserstein & Lazar, 2016).
Note on Methods
All figures were computed from the sample data in Python. Intervals for means use the t distribution; intervals for proportions use the normal approximation where the counts allow it. The regression is an ordinary least squares fit with one predictor, checked by examining residuals by distance and by time of day, following the practice of plotting data rather than relying on summary statistics alone (Anscombe, 1973).
References
Anderson, D. R., Sweeney, D. J., Williams, T. A., Camm, J. D., & Cochran, J. J. (2020). Statistics for business and economics (14th ed.). Cengage.
Anscombe, F. J. (1973). Graphs in statistical analysis. The American Statistician, 27(1), 17-21. https://doi.org/10.1080/00031305.1973.10478966
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
The Herzing BU 321 Unit 8 instructions
The final BU 321 assignment asks for a memo that uses statistical analysis to support one business decision. Most versions want the recommendation stated first, the supporting evidence presented in a form a manager can check, technical detail kept to exhibits, and limits stated plainly. Many rubrics ask you to draw on analyses from earlier units rather than starting over. Charts and tables should be labeled with sources and referred to in the text. The memo should read clearly for someone who never took statistics while remaining checkable by someone who did. Use a memo header and cite any outside sources in APA 7.
Inside the BU 321 Unit 8 example
The sample's memo header is followed by a two-sentence recommendation and a short reason. Exhibit 1 is a table of on-time share and mean delivery time by distance band with its source. Exhibit 2 summarizes the regression and explains why the 35-minute promise fails at the edge of zones. Exhibit 3 shows the chain meets its franchise standard and that the problem is distance rather than one store. A section weighs two alternatives, shrinking zones and adding drivers, and explains why each was rejected. The limits section names four limits, gives an approximate interval for the distant-order on-time share and describes the data that would change the recommendation.
Reading the BU 321 Unit 8 rubric
The rubric for the decision memo generally weighs the clarity of the recommendation, the quality and presentation of evidence, integration of earlier analyses and the treatment of limits. Clarity earns full points when the action appears first in plain words. Evidence is judged by labeled exhibits with sources that directly support the recommendation. Integration rewards memos that reuse interval, test and regression results appropriately rather than repeating them in full. The limits criterion is often decisive and rewards memos that state what the data cannot settle, what would change the decision and what it would cost to find out, as the sample does. Readability for a nontechnical manager and correct citations complete the grade.
BU 321 Unit 8 help: mistakes that cost points
The most common Unit 8 mistake is writing a statistics report with a recommendation added at the end. Another is filling the memo with technical terms that a manager cannot follow. Students often include every analysis from the term instead of those that support the decision. Exhibits sometimes appear without labels or sources, or are never mentioned in the text. Limits are frequently omitted, which makes the recommendation look more certain than the data allow. Finally, consider alternatives; explaining why a simpler option, such as shrinking the delivery zone, was rejected strengthens the case for the one you recommend. Keep the recommendation to one action, since a memo that asks for several changes invites the reader to approve none of them. Test the memo on someone outside the course before submitting.
Get a BU 321 Unit 8 example written to your instructions
Send the Unit 8 instructions and the rubric from your BU 321 classroom, with the dataset and the decision the assignment sets. We write a custom example that opens on the recommendation, keeps notation in numbered exhibits, states the limits after the evidence and names what would reverse the answer. First custom sample free, back in 24 to 48 hours.
More BU 321 papers and business management samples
- BU 321 Unit 1: Descriptive summary set
- BU 321 Unit 2: Distribution problem set
- BU 321 Unit 3: Sampling and error write-up
- BU 321 Unit 4: Confidence interval report
- BU 321 Unit 5: Hypothesis test write-up
- BU 321 Unit 6: Two sample comparison paper
- BU 321 Unit 7: Regression output interpretation
- BU 106 Unit 8: Company analysis with recommendation
- BU 217 Unit 8: Recommendation with defense
- BU 231 Unit 8: Market entry recommendation
- BU 242 Unit 8: Improvement recommendation
BU 321 Unit 8 questions, answered
How much statistics belongs in the memo itself?
As little as the argument allows and as much as the exhibits need. A manager reading the first paragraph should not meet a symbol, while a reader checking the work should find everything they need below it. Where the instructions ask for the analysis in the body, keep the plain reading above it so both audiences are served.
What if the analysis does not support a clear action?
Say that, and recommend the smallest useful step instead: a trial, a longer collection period, a second measurement. A memo manufacturing confidence out of ambiguous data is worse than one recommending a way to resolve the ambiguity, and sections that grade judgment tend to reward the second.
Does the memo need a chart?
Only if a chart shows something the table cannot. One well chosen exhibit displaying the difference or the trend can carry the argument, while a set of decorative charts pushes the recommendation down the page. Follow any requirement in the instructions, and label whatever you include so it can be read without the surrounding text.
Where can I find a free BU 321 Unit 8 sample paper?
The whole BU 321 Unit 8 decision memo, with all three exhibits and the limits section, is free to read above. If your final memo addresses a different decision, share the prompt and data, and you will receive a free custom memo built on them, usually within two days.
How much statistics belongs in the BU 321 Unit 8 memo?
Enough to be checked, kept in labeled exhibits. The sample states its recommendation and reason in plain words and places the intervals, test and regression in exhibits below. A manager can act on the first section, while an analyst can verify the rest.