BUS 655 Topic 6 renewal risk model example

Sports Business Analytics Grand Canyon University Free custom sample in 24 to 48h

This BUS 655 Topic 6 sample paper builds a season member renewal risk model for a composite soccer club and uses it to choose who gets a renewal call. It belongs to GCU BUS 655, Sports Business Analytics, listed as BUS-655 in the Grand Canyon University MBA with an Emphasis in Sports Business program. Lapse is defined as full cancellation, with downgrades tracked apart. Drivers come from the club's own records, led by the share of tickets scanned. On a held-back renewal cycle the top two risk deciles hold 47 percent of cancellations. An accidental split two years earlier shows calls help members whose decision is still open, so the 900-call list favors them.

CourseBUS 655 Sports Business Analytics
TopicTopic 6
Paper typeRenewal risk model
LengthAbout 1,030 words, 4 pages
FormatAPA 7 student paper
SchoolGrand Canyon University (GCU)
ProgramMBA
UpdatedOctober 2026

Free sample paper for BUS 655 Topic 6

1

Who Will Not Renew, and Who Can Still Be Persuaded: A Season Member Renewal Risk Model for a Composite Soccer Club, Ranked by Decile on a Held-Back Season

[Author Name]

Colangelo College of Business, Grand Canyon University

BUS-655: Sports Business Analytics

Topic 6 Assignment

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

What this page is doingThe title separates the two questions the model answers, risk and persuadability, and states how it was tested. The club, its members and every figure are an illustrative composite; the sources are real.
2

Who Will Not Renew, and Who Can Still Be Persuaded: A Season Member Renewal Risk Model for a Composite Soccer Club, Ranked by Decile on a Held-Back Season

What Counts as Lapsing

Riverbend FC has about 2,900 season members, and each spring its six-person ticket office runs a renewal campaign with time for roughly 900 personal calls. The question register asked which members are likely not to renew, so the calls go where they matter. Before modeling anything, the outcome has to be defined, because lapse can mean three things with very different costs. A member can cancel entirely, move to a smaller plan or switch to cheaper seats. Full cancellation loses all of a member's revenue and is the outcome modeled here. Downgrades and seat switches are tracked separately and reported beside the model, since they need a different conversation. All figures below are illustrative composites.

What this page is doingDefining the outcome first prevents the model from mixing a cancellation with a seat change, two events that call for different responses from the ticket office.
3

Drivers From the Club's Own Records

Every driver comes from records the club already holds, joined by account number across the ticketing system and the CRM.

1. Scan share: the share of the member's tickets scanned at the gate last season. A member who paid for 17 matches and came to five is telling the club something.

2. Transfers and unused seats: how often the member forwarded tickets to someone else, and how often seats went unused without being transferred.

3. Tenure: seasons as a member.

4. Plan size: number of seats on the account.

5. Payment method: automatic installments or a single payment.

6. Service complaints: tickets logged in the CRM in the past season.

7. Price change: the percentage change in the member's price at their last renewal.

The Model

Table 1

Estimated Effects on the Odds of Canceling, Illustrative

DriverEffect on odds of cancelingRange
Scan share below 40% (against 70% or more)3.4 times higher2.6 to 4.4
Scan share 40% to 70%1.7 times higher1.3 to 2.2
Five or more unused, untransferred seats1.9 times higher1.5 to 2.4
First or second season as a member1.6 times higher1.3 to 2.0
Single payment (against installments)1.4 times higher1.1 to 1.8
Any service complaint1.5 times higher1.1 to 2.0
Each 5 points of last price increase1.2 times higher1.0 to 1.4

A logistic regression was fitted on the two previous renewal cycles, 5,600 member-seasons, with cancellation as the outcome. Logistic regression was chosen because it gives each driver an effect the ticket office can understand and because it performs well on this kind of problem when the drivers are sensible (Hosmer et al., 2013). Table 1 shows how each driver changes the odds of canceling, holding the others steady.

Attendance is by far the strongest signal. Members who stop using their seats stop paying for them. Unused seats that were not even passed to a friend add to that, while new members and those who faced a larger price increase are more likely to leave.

How Well It Ranks Members

Table 2

Cancellation by Predicted Risk Decile, Held-Back Renewal Cycle

Risk decileMembersActual cancellation rateShare of all cancellations
1 (highest)29046%28%
229031%19%
329024%15%
429018%11%
529014%9%
629011%7%
72908%5%
82906%4%
92904%2%
10 (lowest)2902%1%

A renewal model is useful if it puts the members most likely to cancel at the top of the list, so it was judged on ranking rather than on overall accuracy. The model was applied to members before last spring's renewal, a cycle it had not seen, and members were split into ten equal groups by predicted risk. Table 2 shows what actually happened. Ranking performance on a held-back period is the test that matters for churn models used to target outreach (Neslin et al., 2006).

The top two deciles held 20 percent of members but 47 percent of cancellations. The overall cancellation rate was about 16 percent. The model ranks well enough to guide the campaign.

What this page is doingThe decile table on a cycle the model never saw shows the result the ticket office cares about: how many of the departures sit at the top of its list.
4

Risk Is Not the Same as Persuadability

The obvious use of the model is to call the 900 highest-risk members. That would be a mistake. Some high-risk members have already decided: they moved away, lost interest or attended twice all season. A call will not change their minds. Targeting the customers most likely to leave can waste retention effort, because the customers most likely to respond to outreach are often not the riskiest (Ascarza, 2018).

The club has evidence on this point from an accident. Two springs ago, short staffing meant the ticket office called only members with surnames from A to M. Comparing them with members from N to Z, who were otherwise similar, shows what a call did. Among members with very low scan share, below 20 percent, calls made no measurable difference: about 52 percent canceled either way. Among members with moderate risk and a scan share between 40 and 70 percent, a call lowered cancellations from about 24 percent to about 15 percent.

The Call List

The recommended call list therefore favors members whose risk is meaningful and whose decision is still open. It takes members in deciles 2 to 5 with scan share above 40 percent first, about 700 members, then adds first-season members in decile 1 who attended at least a third of matches, about 200. Members with very low attendance receive an email with a renewal offer and a short survey asking why they did not come, rather than a call.

If the call effect seen two springs ago holds, calling 900 members chosen this way would prevent roughly 75 cancellations, against about 25 if the calls went to the highest-risk members by rank. At an average membership of $480, that difference is worth about $24,000 a season.

Downgrades and Limits

About 6 percent of members downgraded last spring, mostly by dropping one seat from a four-seat plan. These are tracked in a separate report, because the right response is a conversation about plan size rather than a save call. The model's limits are also stated plainly: the call effect rests on one natural split two seasons ago, and the campaign this spring should keep a random 10 percent of the call list uncalled so the effect can be measured again.

References

Ascarza, E. (2018). Retention futility: Targeting high-risk customers might be ineffective. Journal of Marketing Research, 55(1), 80-98. https://doi.org/10.1509/jmr.16.0163

Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied logistic regression (3rd ed.). Wiley.

Neslin, S. A., Gupta, S., Kamakura, W., Lu, J., & Mason, C. H. (2006). Defection detection: Measuring and understanding the predictive accuracy of customer churn models. Journal of Marketing Research, 43(2), 204-211. https://doi.org/10.1509/jmkr.43.2.204

The BUS-655 Topic 6 assignment instructions

The Topic 6 instructions for BUS-655 ask students to model which customers are at risk of not renewing. Weak papers fit a model and recommend calling the riskiest customers. The assignment rewards a clear definition of lapse, drivers drawn from the organization's own records, effects reported with ranges and a test on a period the model did not see, judged by how well it ranks customers. Students should then ask a second question, who can still be persuaded, and design outreach around it. Stating limits and building in a way to measure the effect of outreach next time completes graduate work.

How the BUS-655 Topic 6 example is put together

The paper defines lapse as full cancellation and explains why downgrades are reported separately. Seven drivers are listed, all joined from the ticketing system and CRM. Table 1 shows each driver's effect on the odds of canceling with a range. Table 2 ranks members by risk decile on the held-back renewal cycle. A section explains why the riskiest members are not the best to call, using a past A to M calling split as evidence. The call list then favors moderate-risk members with reasonable attendance, the expected gain is costed and a final section covers downgrades and a random uncalled group.

BUS 655 Topic 6 rubric: what Excellent looks like

Faculty grading the renewal model reward a precise definition of the outcome before any modeling. Strong papers draw drivers from records the club holds and report their effects with ranges. Credit goes to a ranking test on a held-back period, shown by decile, rather than a single accuracy figure. The rubric values the distinction between risk and persuadability, supported by research and by the club's own evidence. The highest band designs the outreach list around that distinction, estimates its value and keeps a random uncalled group so the effect can be measured again. Models that send calls straight down the risk ranking usually lose marks.

Common BUS 655 Topic 6 mistakes, and how to avoid them

The deduction graders reach for first is a model that predicts lapses and never says who gets called. An accurate score with no outreach plan attached answers a question the service team did not ask. Drivers that leak the outcome cost heavily, such as a flag for members who already declined the renewal offer, which produces excellent accuracy and no warning at all. Papers judging the model by overall accuracy miss that most members renew, so a model predicting universal renewal scores well and helps no one; ranking by decile is the test that matters. Treating every lapse as one outcome blurs a lost account with a member who moved to a smaller plan. Finally, a call list longer than the staff can work, or aimed at members certain to leave, spends effort where it cannot change the result.

Get a BUS-655 Topic 6 example written to your instructions

Send the BUS-655 Topic 6 instructions and the rubric in your classroom, plus any account data or case your section is using. A custom example is written to those criteria, with lapse defined, drivers limited to what was known in time, the ranking validated and a call list sized to staff hours, in 24 to 48 hours. The first one is free.

More BUS 655 papers and MBA samples

BUS 655 Topic 6 questions, answered

Why not judge a retention model by its accuracy?

Because when most members renew, a model that simply predicts everyone renews looks accurate and is useless. What the club needs is ranking: are the members scored highest actually more likely to leave? A table of lapse rates by risk decile on a held-back season answers that directly, and it translates straight into how far down the list a call campaign should go.

What data predicts season ticket cancellations?

Usage is typically the strongest signal, since members whose seats go unscanned or are routinely passed on have less reason to renew. Tenure, plan size, payment method, service complaints and the price change at the last renewal often add to it. Team results matter too, but the club cannot act on them, so the model is most useful for the signals outreach can address.

Does a high risk score mean a call will save the member?

Not necessarily. The score estimates who is likely to leave, not who will respond to contact. A member moving out of the region will leave regardless, while one frustrated by unused seats might stay with an exchange option. Testing outreach on part of the list and comparing renewal rates is how a club learns whom a call actually helps.

Where can I find a free BUS 655 Topic 6 sample paper?

The soccer club's renewal model is complete above, from the lapse definition and drivers to both tables, the persuadability evidence, the call list and the limits. Describe the member records your organization keeps and we will draft a free model around them.

Why not call the highest-risk members in BUS 655 Topic 6?

Many of them have already decided. At the club, calls made no difference for members who attended under 20 percent of matches but cut cancellations from about 24 percent to 15 percent among moderate-risk members.