Can you describe a project that didn't succeed?

Certainly! I’d be glad to share an experience regarding a project that didn’t go as planned during my time at Expedia.

Project Description: The project aimed to develop an automated recommendation engine for hotel bookings tailored to user preferences. The objective was to increase conversion rates by 15% by providing more personalized suggestions to users, thus enhancing the overall customer experience.

Challenges Faced:

1. Data Quality Issues: One of the major challenges was the discrepancy and inconsistency in the customer data we had. The data was not as clean or as comprehensive as required for training an effective machine learning model.
2. Integration Complexities: Integrating the recommendation engine into the existing Expedia platform was more complicated than anticipated. We faced several technical hurdles in ensuring smooth interoperability between existing systems and our new solution.
3. Time Constraints: Due to tight deadlines, there was limited time for iterative testing and refinement, which is crucial for the success of such data-driven projects.

Reasons for Failure:

1. Inadequate Data Preparation: The data cleaning and preparation phase was underestimated. As a result, the machine learning model was unable to accurately predict user preferences and the recommendations were not significantly better than the existing solution.
2. Technical Debt and Resource Allocation: The existing technical infrastructure was not fully compatible with our new system, requiring more resources than initially planned. This redirected focus away from optimizing the recommendation engine itself.
3. Scope Creep: During the project, additional features and functions were requested that stretched our resources thin and deviated focus from the primary objectives.

Lessons Learned:

1. Importance of Data Quality: One key lesson learned was the critical importance of having high-quality, well-prepared data. Future projects should allocate appropriate time and resources to data auditing and cleaning before model development begins.
2. Realistic Planning and Communication: We learned to set more realistic timelines and foster transparent communication across teams. Early identification of potential integration issues could have helped in adjusting goals and timelines suitably.
3. Flexible Project Scope: Maintaining a flexible project scope that can adapt based on real-time challenges is essential. It’s important to prioritize core functionalities over additional features to ensure the primary objectives are met.

Improvements Made: Post-project, we initiated a thorough review of our data collection and integration processes. This led to the implementation of more robust data management practices and clearer guidelines on project scope management, ensuring future projects would be better prepared for similar challenges. Additionally, we established more structured cross-functional communication protocols to better manage expectations and resource allocation.

This experience, while not successful in terms of achieving initial objectives, proved invaluable in refining our processes and approach to future projects at Expedia.