What needed to be understood.
The review dataset mixed structured ratings and recommendation fields with open-text comments, prices, age groups, regions, and use cases. The analysis needed to turn those signals into a clear pricing recommendation without overstating what the reviews could prove.
How the analysis was built.
- Structured 270 reviews into analyzable fields for price, rating, recommendation, age group, region, comments, and use case.
- Built Power BI views for aggregate sentiment, rating distributions, price comparisons, demographic patterns, and regional performance.
- Used word clouds to surface recurring language around value, perceived benefits, and product use.
- Translated the dashboard into pricing and targeting recommendations.
What the evidence showed.
- The $10 option generated the strongest overall response and the highest average rating in the submitted analysis.
- The $12 option produced overwhelmingly negative recommendations compared with the lower prices.
- Older customers responded more positively than several younger segments.
- Regional differences existed, but price appeared more influential than geography in the submitted analysis.
What should happen next.
- Use the $10 point as the primary commercial test and treat $11.25 as a controlled secondary experiment.
- Avoid scaling the $12 offer without a stronger value proposition or supporting evidence.
- Test messaging by age segment and track repeat purchase behavior rather than relying only on stated recommendation.
What this project does—and does not—prove.
The source reviews were a structured academic dataset. Recommendation labels and ratings provide directional evidence, but they do not establish product efficacy or guarantee real-market demand.