| Course | MBA FPX 5008 Applied Business Analytics |
|---|---|
| Assessment | Assessment 1 |
| Paper type | Dataset analysis and visuals |
| Length | About 1,080 words, 5 pages |
| Format | APA 7 student paper |
| School | Capella University |
| Program | MBA |
| Updated | September 2026 |
Free sample paper for MBA FPX 5008 Assessment 1
Describing and Analyzing Customer Churn Data for a Regional Internet Service Provider
Student Name
Capella University
MBA FPX 5008: Applied Business Analytics
Instructor Name
Month Day, Year
Introduction
Customer churn, the rate at which customers cancel service, is a central concern for subscription businesses because acquiring a new customer usually costs more than keeping an existing one. This paper analyzes a dataset of customer accounts from a regional internet service provider with about 85,000 residential customers, to understand which customers are most likely to cancel. It describes the data before analysis, declares how the data were cleaned, chooses charts for the questions asked, and states findings at the strength the data allow. The provider is fictional and its figures are illustrative.
The Business Question
The marketing director asked a specific question: which identifiable groups of customers cancel at the highest rates, so that retention offers can be targeted? The analysis is designed to answer that question, not to build a predictive model, which will be the subject of a later assessment.
The Data Described Before Analyzed
The dataset contains one row for each of 12,400 customer accounts sampled randomly from accounts active at the start of last year. Each row includes the customer's tenure in months, plan type (basic, standard, or premium), contract type (month-to-month or 12-month), monthly charge, whether the customer bundles television service, the number of technical support calls in the past six months, whether the customer uses autopay, the region (urban, suburban, or rural), and whether the customer canceled during the year. Tidy data principles, in which each variable is a column and each observation a row, guided the structure of the dataset (Wickham, 2014). Overall, 1,860 of the 12,400 customers canceled during the year, a churn rate of 15.0%. Median tenure was 26 months, and the median monthly charge was 74 dollars. Month-to-month contracts accounted for 58% of accounts.
Cleaning Decisions Declared
Every cleaning decision is declared. Thirty-eight accounts with a monthly charge of zero were identified as employee or promotional accounts and removed, leaving 12,362. Tenure values above 240 months, which predated the company's founding, were found in 12 accounts and traced to a data migration error; they were recoded as missing and excluded from tenure analyses only. Support calls were missing for 410 accounts because of a gap in the call center system during one month; these accounts were kept, and support call analyses were run on the remaining accounts. Region was standardized from 14 variant spellings to three categories. Declaring cleaning decisions allows another analyst to reproduce the dataset and judge whether the choices could have changed the results.
Charts Chosen for the Question
Charts were chosen to answer the business question clearly. A horizontal bar chart shows churn rates by contract type, plan type, region, and autopay use, sorted from highest to lowest, because the question is which groups churn most. A line chart shows churn rate by tenure band, because churn is expected to change with tenure. A bar chart shows churn by number of support calls. Pie charts were avoided because comparing slices is harder than comparing bar lengths. Tufte (2001) argued that good graphics show the data clearly and avoid decoration that does not carry information, a principle applied here.
Findings
Churn differed sharply by contract type: 22.6% for month-to-month customers and 4.5% for 12-month contract customers. Churn was highest in the first year of tenure, at 27%, and fell steadily to 8% for customers with more than four years. Customers not using autopay churned at 19.8%, compared with 10.1% for autopay users. Customers with three or more support calls in six months churned at 31%, compared with 12% for those with none. Rural customers churned at 18.4%, compared with 13.9% for urban customers. Plan type and bundling showed smaller differences: basic plan customers churned at 16.8% and premium plan customers at 12.9%, while bundled and unbundled customers differed by less than two percentage points.
Findings Stated at the Strength the Data Allow
These are associations, not causes. Month-to-month customers may churn more because the contract allows it, or because customers who expect to leave choose month-to-month. Support calls may cause dissatisfaction, or dissatisfied customers may call more. The analysis identifies groups to target, which answers the marketing director's question, but it does not show that changing contract type or autopay status would reduce churn. The overlap among groups is also important: many new customers are on month-to-month contracts without autopay, so the high churn in each group partly reflects the same customers.
A Closer Look at Tenure and Contract Together
Because new customers and month-to-month customers overlap, a combined view was prepared. Among customers in their first year, month-to-month customers churned at 34%, while those on 12-month contracts churned at 7%. Among customers with more than two years of tenure, the gap narrowed to 11% versus 3%. The pattern suggests that the first year is the critical window and that contract type matters most during that window. Looking at the two variables together shows where the risk is concentrated in a way that separate bar charts could not.
Data Quality Beyond Cleaning
Some data quality issues could not be fixed by cleaning. The support call variable counts calls but not their outcome, so a customer whose problem was solved in one call looks the same as one whose problem was never solved. The region variable is based on billing address, which may differ from service address for a small number of accounts. These limits are noted so that the marketing team does not overinterpret small differences.
What the Findings Suggest for Targeting
The findings suggest that retention efforts should focus on customers in their first year, on month-to-month contracts, without autopay, and with multiple support calls. A cross-tabulation showed that the 1,240 customers meeting three or more of these conditions churned at 38%.
Limitations
The data cover one year and one provider. They do not include customer satisfaction, competitor offers, or reasons for cancellation, which would help explain churn. A short exit survey for canceling customers, and a matching of accounts to areas where competitors recently launched fiber service, would add the explanatory variables this dataset lacks and would strengthen the modeling planned for the next assessment.
Conclusion
Describing the data before analysis, declaring cleaning decisions, choosing charts for the question, and stating findings at their true strength produced a clear answer: churn is concentrated among new, month-to-month customers without autopay and with repeated support calls.
References
Tufte, E. R. (2001). The visual display of quantitative information (2nd ed.). Graphics Press.
Wickham, H. (2014). Tidy data. Journal of Statistical Software, 59(10), 1-23. https://doi.org/10.18637/jss.v059.i10
Reading the MBA-FPX5008 Assessment 1 instructions
The first assessment in MBA FPX 5008 asks you to analyze a business dataset and present the results with visuals. You state the question the analysis answers, describe the dataset, including its source, size and variables, explain how you cleaned it, such as handling missing values and duplicates, calculate descriptive statistics, create charts suited to each comparison, interpret the findings without overstating them and note data quality issues. Churn, the share of customers who leave in a period, is a common business measure for this kind of analysis. Business analytics texts and the software documentation for Excel or similar tools support the work.
How this MBA FPX 5008 Assessment 1 example is built
Data, preparation, visuals, findings. The opening describes the dataset and how it came to exist, including its limitations. A preparation block states every decision taken before analysis: what was cleaned, what was excluded and why, and what that might do to the result. A visuals block presents the charts, each chosen for the question it answers, with axes and captions that let a reader interpret it unaided. A findings block describes what the data shows, distinguishing what is clear from what is suggestive. A limits block names what the dataset cannot answer, which at graduate level carries real credit. The closing states what further data would settle the open questions. Nothing is claimed as causal that the design cannot support.
Where the marks sit in the MBA FPX 5008 Assessment 1 scoring guide
Capella's scoring guide rates the business question, the data description, the cleaning, the statistics, the visuals and the interpretation. The question reaches Distinguished when it names a decision the analysis will inform, such as where to focus retention offers. The data description scores higher when it covers source, size, period and variables before any analysis. Cleaning earns credit when each decision is declared, like removing test accounts. Statistics are judged on accuracy and relevance. Visuals should be chosen for the comparison, such as a bar chart of churn by contract type. Interpretation should stay within what the data show, since month-to-month customers churning more is an association, not proof that contracts cause loyalty. Data quality should be noted. Proficient analyses describe the data correctly; Distinguished ones answer the business question.
MBA FPX 5008 Assessment 1 help from the desk
Dataset analyses often produce every possible chart without asking which ones answer the question. Another weakness is skipping the data description, so readers do not know what the numbers represent. Cleaning decisions are made silently, which makes the results impossible to reproduce. Pie charts are used for comparisons they cannot show well. Findings are stated as causes, like "short contracts cause churn." Data limits, such as missing reasons for cancellation, are left out. Stating the question first, describing the data in a short table and then choosing two or three charts that answer the question keeps the analysis focused and prepares the costed recommendation in Assessment 2.
Get a MBA FPX 5008 Assessment 1 example written to your instructions
Send the Assessment 1 instructions and the scoring guide from your MBA FPX 5008 courseroom, plus the dataset your version supplies. We write a custom example against those exact criteria and return it in 24 to 48 hours. The first custom sample is free, and analytics answers come entirely from the data given, so send yours. The paper above is an original model document written by our desk, not a submitted student paper and not an official Capella University document.
More MBA FPX 5008 papers and MBA samples
- MBA FPX 5008 Assessment 2: Costed Recommendation from Data
- MBA FPX 5008 Assessment 3: Analytics Synthesis
- MBA FPX 5010 Assessment 1: Statement Reading for Managers
- MBA FPX 5006 Assessment 1: Strategic Process and Analysis
- MBA FPX 5012 Assessment 1: Segmentation and Targeting
- MBA FPX 5002 Assessment 1: Leadership Self-Assessment
MBA FPX 5008 Assessment 1 questions, answered
How much cleaning should I do?
As little as the analysis requires, and state all of it. Removing impossible values is defensible; removing inconvenient ones is not, and the difference is whether you can explain why a record could not be real. Report how many records you dropped and what the result looks like with them included.
Which chart should I use?
Whichever makes the comparison the question needs. Distributions want a histogram, relationships want a scatter, change over time wants a line. The test is whether somebody could read the answer off the chart without your paragraph beneath it, and if they cannot, the chart is decorative.
Can I say one thing caused another?
Only if the data was collected in a way that supports it, which observational data usually is not. Describe what the relationship is and what else could explain it. Naming a plausible confounder yourself is one of the fastest ways to demonstrate the analytical care these criteria reward.
Where can I find a free MBA FPX 5008 Assessment 1 sample paper?
The internet provider churn analysis is available on this page in full, with the business question stated, the data described before analysis, cleaning decisions declared, charts chosen for the question and findings at the strength the data allow.
Can I say one thing caused another in MBA FPX 5008 Assessment 1?
Not from descriptive data alone. You can report that two things move together, but showing cause needs an experiment or a design that rules out other explanations.