Background

A small inn was exploring options to expand its recreational offerings by adding horseback riding or ice skating to attract more guests and increase revenue. However, with no prior experience offering these services and limited historical data, the inn faced significant uncertainty about which option, if any, would be financially viable. Making the wrong investment could lead to wasted capital and missed opportunities, so a rigorous, data-driven approach was essential to guide this strategic decision.

Challenges

This project faced several unique challenges from the outset. Primarily, the inn’s clientele skewed older, meaning that traditional survey methods—especially digital surveys—risked low engagement and incomplete responses. Older guests were less likely to fill out lengthy questionnaires, and many had limited familiarity or comfort with online forms. This made collecting reliable, high-quality preference data a major hurdle.

To overcome this, I designed a carefully crafted survey in Qualtrics that balanced comprehensiveness with accessibility. The survey employed clear, concise language and user-friendly formats to minimize respondent fatigue, while still capturing the detailed trade-offs needed for discrete choice modeling. Crafting questions that simulated real-world decisions—such as choosing between horseback riding, ice skating, or no additional service—was key to understanding how guests valued each option and what factors influenced their choices.

From a modeling perspective, discrete choice models require robust data reflecting real preferences, but since these services were new to the inn, there was no historical usage data to lean on. This increased the uncertainty and complexity of accurately forecasting demand. The model needed to translate survey responses into probabilities that guests would select each option, taking into account various attributes and constraints.

Integrating these behavioral insights with financial data—such as the costs of implementing each service and projected revenues—added another layer of complexity. Close collaboration with the inn’s management ensured that all assumptions were realistic and grounded in operational realities. Balancing statistical rigor with practical business knowledge was critical to building a trustworthy model.

Solution

To address these challenges, I developed a discrete choice model that leveraged the survey data to estimate the likelihood of guests choosing horseback riding, ice skating, or no new service. The survey was meticulously designed in Qualtrics to feed directly into the model, capturing the trade-offs guests make when evaluating recreational options.

Using Python, I implemented the discrete choice framework to analyze customer preferences, quantify demand, and simulate how different factors influenced guest decisions. This approach models decision-making as a function of multiple attributes, such as cost, convenience, and personal interest, reflecting the nuanced ways customers weigh their options.

The model then combined predicted demand with financial projections—incorporating costs of setup, maintenance, and expected revenues—to forecast overall profitability for each service option. Finally, I visualized the results in Tableau to present clear, actionable insights to the inn’s leadership.

Impact

This analysis delivered critical insights that saved the inn tens of thousands of dollars by preventing an unprofitable expansion. The analysis revealed that neither horseback riding nor ice skating would generate a positive return on investment given the projected demand and costs.

Beyond the immediate financial impact, the project showcased the power of combining well-designed surveys, behavioral modeling, and visualization to guide strategic decisions—especially in contexts with limited historical data and uncertain demand. By quantifying guest preferences and linking them to profitability, the inn’s leadership gained confidence in their growth strategy while avoiding costly missteps.

This project also highlighted the importance of tailoring data collection and modeling approaches to the specific characteristics of the customer base, ensuring insights are both accurate and actionable.