Marzooq

Chapter · 01

Research Questions

Five questions structure the inquiry — from identifying drivers of satisfaction to deploying predictions as concrete operational strategy.

RQ1

Key Influencing Factors

Which operational, product, and customer attributes have the strongest measurable impact on customer review ratings?

How this helps the company

By identifying which operational, product, and customer‑related factors most strongly influence review outcomes, the company gains a clear understanding of what truly matters to customers. This allows managers to prioritize improvements in areas such as delivery performance, packaging quality, and pricing transparency. Instead of guessing, the company can make decisions based on evidence.

RQ2

Predictive Capability

Can machine learning models accurately predict customer review outcomes using the available business data?

How this helps the company

The machine‑learning models developed in this project allow the company to predict which orders are likely to receive negative reviews. This enables proactive intervention — contacting the customer, expediting delivery, or offering support before dissatisfaction escalates. This reduces the number of negative reviews and protects the company's reputation.

RQ3

Best Model Selection

Which machine learning model (Logistic Regression, Random Forest, or XGBoost) provides the most reliable performance for business‑oriented review prediction tasks?

How this helps the company

By comparing Logistic Regression, Random Forest, and XGBoost, the project identifies the model that provides the most reliable performance. This helps the company avoid wasting time and resources on ineffective approaches and ensures that the chosen model can be deployed confidently in real‑world operations.

RQ4

Feature Contribution

How do individual features (such as delivery time, product attributes, or pricing) contribute to the prediction of customer review outcomes, and which variables hold the highest business relevance?

How this helps the company

Feature importance analysis reveals which variables have the greatest impact on customer satisfaction. This helps the company focus its resources on the areas that will produce the biggest improvements. For example, if delivery delays are the strongest predictor of negative reviews, the company can invest in logistics optimization rather than unrelated areas.

RQ5

Business Impact

How can the insights from machine learning predictions be translated into actionable business strategies that improve customer satisfaction and operational performance?

How this helps the company

The final research question ensures that the project does not stop at analysis. Instead, it translates the findings into clear, data‑driven business strategies, such as: improving delivery reliability; enhancing packaging for fragile or heavy items; strengthening customer communication; monitoring payment‑related friction points; and using predictive analytics for early issue detection.

Outcome

From questions to strategy.

Together, these five research questions form a coherent arc: identify the variables that move review outcomes, validate that machine learning can exploit them at scale, select the most reliable algorithm, surface the drivers in an interpretable way, and finally — translate those drivers into operational decisions.

  • Improving delivery reliability
  • Enhancing packaging for fragile or heavy items
  • Strengthening customer communication
  • Monitoring payment‑related friction points
  • Using predictive analytics for early issue detection

These strategies help the company improve customer experience, reduce operational inefficiencies, and increase long‑term customer loyalty.