Reducing Idle Time and Improving Revenue Visibility

Executive Summary

This project focused on identifying operational inefficiencies within a ride-sharing system, specifically around driver utilisation and revenue performance.

By restructuring raw trip data and analysing patterns across time and location, I uncovered clusters of driver idle time and mismatches between supply and demand. The analysis showed that high trip volume did not necessarily translate to efficient operations.

Based on these insights, I proposed targeted strategies to improve driver allocation, reduce idle time, and stabilise revenue performance.

Business Problem

Ride-sharing platforms operate in high-volume, real-time environments where performance is often measured through surface-level metrics such as total trips or revenue.

However, these metrics can mask underlying inefficiencies.

Key challenges included:

• Drivers spending significant time idle between trips
• Inconsistent revenue performance despite high activity
• Limited visibility into where and when inefficiencies occur

Additionally, fragmented data and inconsistent reporting make it difficult for operational teams to:

• identify underutilised capacity
• respond effectively to demand fluctuations
• optimise resource allocation

The core problem was not a lack of data, but a lack of structured insight into operational efficiency.

Methodology

I approached the problem using backwards induction from business decisions rather than starting with tools or dashboards.

1. Define Key Questions

Focused on operational decision points:

• When and where does idle time occur?
• How does utilisation vary across time and location?
• Where is revenue being impacted by inefficiency?

2. Data Structuring & Transformation

Raw trip-level data was cleaned and transformed into analytical metrics, including:

• idle time between trips
• trip frequency and duration
• revenue per trip and per time window

This created a consistent framework for comparing performance across:

• time periods
• geographic areas
• demand levels

3. Analysis & Visualisation

Instead of building a generic dashboard, I focused on surfacing insights that directly supported decision-making:

• demand vs supply imbalances
• clusters of high idle time
• discrepancies between trip volume and revenue efficiency

The goal was not visual complexity, but clarity and usability.

Skills

• Data analysis and transformation (SQL / Python)
• Data visualisation and dashboard design
• Business problem structuring
• Operational performance analysis
• Translating data insights into business recommendations

Results & Business Recommendation

The analysis revealed several consistent patterns:

• Idle time is concentrated, not random
• Specific time periods showed repeated inefficiencies, suggesting predictable demand-supply mismatches
• High activity does not equal high efficiency
• Some of the busiest periods still underperformed due to poor driver allocation
• Geographic imbalance impacts utilisation
• Certain areas consistently showed lower efficiency despite overall system demand

Business Recommendations

Based on these findings, the following actions would deliver the most immediate impact:

• Pre-position drivers ahead of demand spikes
• Shift from reactive to predictive allocation
• Introduce location-based incentives
• Encourage better distribution of drivers across regions
• Track idle time as a core KPI
• Move beyond volume-based metrics to efficiency-based performance tracking

Impact

These changes would likely result in:

• Reduced driver idle time
• Improved utilisation across time and location
• More stable and efficient revenue generation

More importantly, they enable a shift from:

• measuring activity to measuring efficiency

Next Steps

To extend this work further, I would focus on:

• Demand forecasting
• Predict high-demand periods to improve allocation decisions
• Real-time data integration
• Enable faster operational responses
• Experimentation framework
• Test interventions such as pricing strategies or driver incentives