DAT 565 Week 2 Reports and Descriptive Statistics Example

Averages That Hide the Story: Building Client Performance Reports at an Atlanta Fulfillment Warehouse

[Student Name]

University of Phoenix

DAT/565: Data Analysis and Business Analytics

Week 2 Assignment

[Instructor Name]

[Date]

Peachtree Fulfillment Partners, its clients, data and figures are composites written for a model paper.

Week 1 prepared 12 months of order data for Peachtree, the fictional Georgia warehouse that serves 34 client brands. The table holds one row per order, about 2.5 million in all, with flags for accuracy and on-time shipping and details of each order's lines, units and products. This paper builds reports and descriptive statistics on that table.

Defining the Measures

Neely et al. (1995) reviewed research on performance measurement and argued that measures should be derived from strategy, defined precisely, including formula, frequency and data source, and designed to prompt action. Peachtree's three core measures are defined accordingly. Order accuracy is the share of shipped orders with no client-reported picking or packing error within 30 days. On-time shipping is the share of orders receiving a carrier scan by the cutoff time on the day of release, excluding the carrier outage days identified in Week 1. Time to ship is the hours from release to first carrier scan.

The Overall Picture

Across all clients and months, order accuracy was 99.4 percent and on-time shipping 96.8 percent, against typical contract targets of 99.5 and 97 percent. Mean time to ship was 4.6 hours. These averages suggest a warehouse slightly below target but broadly healthy, which is how the operations team has described itself.

Looking at Distributions

Anscombe (1973) showed with four small data sets that shared nearly identical means, variances and correlations but looked completely different when graphed, arguing that analysts should examine data visually rather than rely on summary statistics alone. Tukey (1977) made the same case for exploratory data analysis, urging analysts to look at data with simple displays before fitting models or reporting summaries. Plotting Peachtree's measures shows why.

Time to ship is strongly right-skewed. The median is 3.1 hours, below the mean of 4.6, and the 95th percentile is 11.4 hours. Most orders move quickly; a tail of orders released late in the afternoon waits for the next wave, and some miss cutoff. The mean blends these groups and describes neither.

Accuracy by client is not normally distributed either. Thirty clients cluster between 99.4 and 99.9 percent; three sit between 97.6 and 98.4 percent; one is at 99.0. The company-wide average hides the three clients who drive most complaints.

Comparing Clients

Grouping by client reveals the pattern behind accuracy. The three low-accuracy clients sell cosmetics and personal care products in many shades and sizes that look alike, and their orders average 4.8 lines against 1.9 for other clients. Their error rate per line is only modestly higher than average; their error rate per order is much higher because each order has more chances for a mistake. That distinction matters: the problem is partly order complexity, not only picker care.

On-time shipping varies more by release time than by client. Clients whose orders arrive mostly after 2 p.m. show on-time rates near 93 percent; those whose orders arrive in the morning exceed 98 percent.

The three clients with the most complaints do not have careless pickers; they have orders with five chances to go wrong instead of two.

Per-Order and Per-Line Measures

The client comparison raises a measurement question. Contracts define accuracy per order, so a client with five-line orders is judged on whether all five lines are right. Per-line accuracy, the share of lines picked and packed correctly, is a fairer measure of picker performance across clients with different order sizes. Across all clients, per-line accuracy was 99.86 percent; for the three low-accuracy clients it was 99.62 percent. Reporting both lets managers separate two questions: how well the warehouse picks, and how much risk each client's order profile creates. The client report will continue to show contract accuracy, since that is what clients pay for, but internal reports will add per-line accuracy.

Release Timing in Detail

Splitting orders by the hour they were released shows a sharp pattern. Orders released before noon shipped on time 99.1 percent of the time, with a median time to ship of 2.2 hours. Orders released between noon and 2 p.m. shipped on time 97.4 percent of the time. Orders released after 2 p.m., about 22 percent of volume, shipped on time only 91.8 percent of the time, with a 95th percentile time to ship of 19 hours, meaning many waited overnight. Six clients send most of their orders in a single afternoon batch from their e-commerce platforms. A conversation about release timing with those clients may improve on-time shipping more than any change inside the warehouse.

Seasonality

Monthly figures show that accuracy dipped in November and December, falling to 99.1 percent overall, when temporary associates make up about 40 percent of pickers. On-time shipping also fell in peak weeks, though less than expected given volumes more than double the normal level. These seasonal patterns mean that annual averages understate peak risk and that comparisons between months must account for staffing mix.

Variability and Outliers

Standard deviations by day show that on-time shipping is most variable on Mondays, when weekend orders pile up, and during promotions. Outlier days, more than three standard deviations below the mean on-time rate, were traced to the carrier outage, two severe storms and one system upgrade. Identifying causes keeps outliers from being either ignored or treated as normal.

The Current Report

The current weekly report is a 14-page spreadsheet listing every client's volume and measures for the week. Managers say they look only at the company totals. It shows no targets, no trends and no exceptions.

A Redesigned Report

The redesigned report has one page. A table lists each client's accuracy and on-time rate for the week and the trailing eight weeks, each compared with its contract target and colored only when below target. A second table shows the median and 95th percentile time to ship by release window. A short exceptions section lists clients below target two weeks in a row with a one-line explanation. Details move to a linked appendix.

Who Receives the Report

The one-page report goes to the operations director, the three shift managers and the account management team every Monday morning. Account managers use it to prepare for client calls, so a client below target hears about the problem and the plan from Peachtree before raising it. The appendix stays available for anyone who needs to investigate a specific client or day.

What the Statistics Suggest for Week 3

The descriptive work points to questions for visualization and deeper analysis: how order complexity relates to errors across all clients, how release timing affects on-time shipping and whether staffing patterns explain Monday variability.

Conclusion

Company-wide averages made Peachtree look nearly on target. Descriptive statistics chosen for each measure's shape, medians and percentiles for time, distributions and group comparisons for accuracy, revealed that a few clients with complex orders and late releases drive most problems. A one-page report built around targets, trends and exceptions gives managers what they need each week.

References

Anscombe, F. J. (1973). Graphs in statistical analysis. American Statistician, 27(1), 17-21. https://doi.org/10.1080/00031305.1973.10478966

Neely, A., Gregory, M., & Platts, K. (1995). Performance measurement system design: A literature review and research agenda. International Journal of Operations & Production Management, 15(4), 80-116. https://doi.org/10.1108/01443579510083622

Tukey, J. W. (1977). Exploratory data analysis. Addison-Wesley.