← Back to all work
Cluster analysis Survey design Course case

FLIP: Auditing a Survey, Segmenting a Market

Critique of a flawed questionnaire, a four-cluster hierarchical segmentation, and segment-specific certification strategies for FLIP's professional development product.

Course Market Research
Method Hierarchical clustering
Sample 50 rows, 18 columns
Segments 4 (Rookies, Seasoned, Veterans, Specialists)
Reading the dendrogram

Reviewing the questionnaire.

Read the questionnaire in Appendix 2 carefully. Like Anand, I have concerns. There are issues with the construction of the questions and the scales used to collect responses.

  • The first two questions are used as filters, which may narrow the data and create biased responses.
  • The questionnaire assumes that only people who are actively pursuing professional certifications should take it, which may overlook the people who may not need those certifications.
  • The questionnaire also has a class ranking question, which may make some respondents uneasy.
  • This is unnecessary on a questionnaire and does not represent everybody with the top 10% or top 25% answer.
    • Having a "neither" option may be unfair for respondents because it represents the entire bottom 75%, which is very broad.
    • They should create more options such as Bottom 50%, top 30%, prefer not to say.
    • And they should also explain why they need this information, as it seems a little irrelevant.
  • Every question should also have a "Not applicable" option.

Cluster analysis.

Running a cluster analysis (without discriminant analysis) on the data, there are four clusters here. This is visible by looking at the dendrogram and segment profiles.

Cluster 1, Rookies (42% of population)

  • Students who are about to or have recently graduated with very limited working experience.
  • They are likely trying to enhance their resumes and will work in finance in the future.

Cluster 2, Seasoned employees (38% of population)

  • People who are in the workforce and have decent work experience.
  • They are focusing on their certifications and trying to get more relevant work in their field.

Cluster 3, Veterans (10% of population)

  • Employees who have a lot of work experience and do not engage with extra activities.

Cluster 4, Specialists (10% of population)

  • Employees that have a lot of BFS experience and prioritize comfortable e-learning.

Re-running with discriminant analysis.

When running a cluster analysis with discriminant analysis, the tool flagged a high-collinearity warning. Several variables (B-school = Crème De La Crème, Major Specialization = Marketing, Minor Specialization = Finance, Minor Specialization = Systems, UG Degree = BCom CA, and Percentile in Class = Top 10 Percentile) were removed because of high collinearity with variables already present in the report.

OptionSelection
Clustering methodHierarchical
Standardization methodNone
Segments forced4
Run discriminant analysisYes
Run classification analysisYes

Next steps for FLIP.

My recommendations for FLIP include strategies to better target each segment.

  • For the rookies, focus on improving their resumes and gaining certifications.
  • For the seasoned employees, focus on certifications that improve current job skills and allow for potential job switching or job improvement.
  • For the veterans, establish more specialization and convenience such as a more flexible schedule and online learning programs.
  • For the specialists, focus on convenience and having self-paced e-learning courses.

I would also collaborate with more B-schools and make FLIP a well-known program. This can help because it makes FLIP certifications credible. Workshops, webinars, and informational sessions can help gain awareness.