| Course | AC570 Data Analytics for Accountants |
|---|---|
| Unit | Unit 9 |
| Paper type | Exception prioritization report |
| Length | About 740 words, 5 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | MS in Accounting |
| Updated | October 2026 |
Free sample paper for AC570 Unit 9
Ranked by Count, Then by Expected Loss: A Unit 9 Exception Prioritization Report for a Storage Chain
[Student Name]
Purdue University Global
AC570: Data Analytics for Accountants
Unit 9 Exception Prioritization Report
[Instructor Name]
[Date]
Stowfield Storage and every figure are composite; the textbook, book and guide cited are real; probabilities are the analyst's stated judgments.
Memorandum
To: Controller and Vice President of Operations From: Data Analyst, Finance Date: February 20, 2026 Re: Which exceptions to act on first
Why Count Is the Wrong Ranking
The exception tests produced seven findings of very different sizes and meanings. Listed by number of items, the after-hours portal credits come first with 1,229, but they are the website granting the move-in deal to customers who sign up after hours: expected, approved and harmless. Ranking by count would send investigators to the least important finding. The report instead ranks by likely cost: the dollars exposed times a stated probability that those dollars represent a loss. The probabilities are judgments, stated so they can be challenged (American Institute of Certified Public Accountants, 2017).
The Ranking
1. Facility 22 cluster. Credits $36,730.50 above the regression's expectation, plus 21 missing receipts valued at the facility's $144.87 average receipt, $3,042.26, for $39,772.76 exposed. Probability of loss 0.60, because four independent tests point to the same facility and no documented cause has been found. Likely cost $23,863.66. Owner: district manager with internal audit. Action: field review within two weeks.
2. Credits above one month's rent without district approval. 312 credits, $61,840 exposed. Probability 0.15, since most are likely legitimate goodwill to long-term tenants but none were approved. Likely cost $9,276. Owner: vice president of operations. Action: approval rule.
3. Unexplained duplicate refunds. 19 pairs, $2,087.13 exposed. Probability 0.90, since both payments cleared. Likely cost $1,878.42. Owner: accounts payable. Action: recovery letters.
4. Late manual credits outside Facility 22. 14 credits, $2,310 exposed. Probability 0.30. Likely cost $693. Owner: district managers. Action: confirm with each user.
5. Missing receipts outside Facility 22. 7 numbers, about $1,014 exposed at average values. Probability 0.40. Likely cost $406. Owner: facility managers.
6. Facility 9's positive residuals. $8,810.65. Probability 0.05, since the seminar showed the pattern reflects lease-up and higher rents. Likely cost $441. Owner: none; add variables to next year's model.
7. After-hours portal credits. 1,229 credits, about $159,000 of promotions. Probability near zero. Likely cost nil. No action beyond confirming the portal's promotion rule.
Ranking by expected loss moves Facility 22 to first by cost and moves the portal credits from first to last, because the question for management is where money is likely being lost, not where the most exceptions happen to be.
Sensitivity
The ranking holds under reasonable changes in the probabilities. Even at a probability of 0.25, Facility 22's likely cost of about $9,943 would still rank first. The approval-rule finding would overtake it only if its probability rose above 0.39, which would mean most large unapproved credits were improper, a conclusion no test supports.
How the Probabilities Were Set
Each probability reflects the evidence behind a finding, not a feeling about the people involved. Findings supported by several independent tests at the same location, like Facility 22, receive higher probabilities than findings supported by one test. Findings with a documented innocent explanation, like the portal credits and most of Facility 9's residuals, receive very low probabilities. Duplicate refunds that cleared twice receive a high probability because the evidence of loss is direct: the money left twice. Where judgment is uncertain, the report errs low, so that the ranking does not overstate the case. A reader who disagrees with any probability can substitute their own and recompute the likely cost; the sensitivity section shows how far each would have to move to change the order.
What Happens to Low-Ranked Findings
Findings ranked fourth to seventh are not ignored. They move into monthly monitoring, where the same tests run on each month's data. If any of them grows, it will rise in the ranking on its own.
The ranking will be refreshed once the Facility 22 review reports, since its findings will change several probabilities at once.
Recommendations
Act on the first three findings now, decide the approval rule at the March 2 meeting and fold the rest into routine monitoring. Expected loss is a screen for attention, not a measure of fraud, and the field review at Facility 22 should begin without assuming wrongdoing (Nigrini, 2012; Richardson et al., 2023).
References
American Institute of Certified Public Accountants. (2017). Guide to audit data analytics. AICPA.
Nigrini, M. J. (2012). Benford's law: Applications for forensic accounting, auditing, and fraud detection. Wiley.
Richardson, V. J., Teeter, R. A., & Terrell, K. L. (2023). Data analytics for accounting (3rd ed.). McGraw Hill.
AC570 Unit 9 instructions, in plain terms
Unit 9 of AC570 asks you to prioritize analytic findings for action. Explain why ranking by count misleads. For each finding, give the number of items, the dollars exposed, a stated probability of loss with its reasoning, the likely cost, an owner and a next action. Assign overlapping items to one finding before summing. Explain how the probabilities were set, tying each to the strength and independence of the evidence, and invite readers to substitute their own. Test the ranking's sensitivity by finding what probability would reorder the top findings. Say what happens to lower-ranked findings. Close with recommendations, making clear that expected loss directs attention and does not establish fraud. Write as a memo to the decision-makers. List the AICPA guide, Nigrini and Richardson and colleagues as APA 7 references.
How this AC570 Unit 9 example is built
The sample explains that 1,229 portal credits lead by count but carry no risk. Seven ranked findings follow: Facility 22's $36,730.50 excess plus 21 missing receipts at $144.87 each, exposed $39,772.76 at 0.60; 312 unapproved credits above a month's rent, $61,840 at 0.15; 19 duplicate refunds, $2,087.13 at 0.90; late manual credits and missing receipts elsewhere; Facility 9's residuals at 0.05; and the portal credits at nil. The highlighted sentence explains how the ranking flips. Sensitivity shows Facility 22 stays first even at 0.25. Sections explain how probabilities were set and what happens to low-ranked items before recommendations. Owners and next actions sit beside every finding so the meeting can assign work at once.
AC570 Unit 9 rubric: what earns full points
Prioritization reports in AC570 earn points for a defensible ranking method and transparent judgments. Graders check that each finding shows items, exposure, probability, likely cost, owner and action, and that overlapping items are counted once. The rubric rewards explaining probabilities in terms of evidence, testing sensitivity and stating clearly that expected loss is a screen rather than a finding of wrongdoing. Recommendations that separate immediate action from monitoring show judgment. Reports that rank by count or dollars alone, assign probabilities without reasons or present the top finding as proven fraud are marked down sharply. Send it as a memo with APA 7 references.
Common AC570 Unit 9 mistakes, and how to avoid them
Weak Unit 9 reports list findings in the order the tests produced them, or by count, so harmless system activity ranks first. Probabilities are assigned without explanation or not at all. Items flagged by more than one test are counted twice. No sensitivity analysis shows whether the ranking is stable. The top-ranked finding is described as fraud. Lower-ranked findings disappear rather than moving to monitoring. Given your own exception register, we can rebuild the ranking, its probabilities and the sensitivity test. Some reports also forget to name an owner for each action, so nothing happens after the meeting.
Get an AC570 Unit 9 example written to your instructions
Collect the exceptions your earlier AC570 units produced, or the list your Unit 9 prompt supplies, and send them with the rubric and any materiality guidance in the case. The ranking, finding briefs and a sensitivity note return in 24-48h with the calculations visible. Nothing is charged for a first custom sample. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More AC570 papers and MS in Accounting samples
- AC570 Unit 1: Data Analytics Question Discussion
- AC570 Unit 2: Analytics Project Scoping Memo
- AC570 Unit 3: Data Preparation Log
- AC570 Unit 4: Data Population Profiling Exercise
- AC570 Unit 5: Full-Population Exception Tests
- AC570 Unit 6: Regression Analysis Exercise
- AC570 Unit 7: Data Visualization Critique
- AC570 Unit 8: Data Interpretation Seminar Reflection
- AC570 Unit 10: Management Presentation
- AC566 Unit 9: Permanent Establishment Treaty Analysis
- AC501 Unit 9: VIE Disclosure Note Draft
- AC557 Unit 9: Control Remediation Plan
- AC502 Unit 9: Auditor Liability Case Study
AC570 Unit 9 questions, answered
Where do the probabilities of loss come from?
From stated judgment, argued finding by finding, since no dataset supplies them. The sample ties each to evidence: three independent tests converging justify a high figure, while a known system behavior justifies zero. What graders look for in AC570 is that the reasoning is visible and the ranking is tested against it. A sensitivity note showing where the order would change turns an assumption into something a reader can check.
How should overlapping findings be handled?
Assign each flagged item to one finding before summing, and say which rule decided it. In the sample, any Facility 22 item joins the facility cluster, whatever test flagged it, because the response there is a single field review. Items elsewhere stay with their own test. Without that step the same dollars inflate several rows, and the ranking overstates the problem it is meant to size.
Where can I find a free AC570 Unit 9 sample paper?
This page holds a full AC570 Unit 9 exception prioritization report for Stowfield Storage ranking seven findings by expected loss with probabilities, sensitivity and APA 7 references.
Why is Facility 22's probability set at 0.60?
Four independent tests, duplicate refunds, missing receipts, late manual credits and amounts just under an approval threshold, point to the same facility, and no documented cause has been found. Converging evidence justifies a higher probability.
How stable is the ranking?
Facility 22 stays first even at a probability of 0.25. The approval-rule finding would overtake it only above 0.39, which would mean most unapproved large credits were improper, a conclusion no test supports.