Summary
After a multibillion-dollar logistics provider turned to Evalueserve to clean and harmonize their data, they wanted to use AI to drive go-to-market activities, including trailer forecasting and customer segmentation. Our trailer forecasting gave the client a 3.8% year-over-year improvement on Linehaul On-Time delivery, which led to a ~$8-14 million impact on operating income. Evalueserve’s domain-specific AI work in customer segmentation, churn, and lifetime value (LTV) resulted in a ~$1 million improvement in operating income and identified another ~$52 million in opportunities.
Background
In 2016, a multibillion-dollar logistics provider began working with Evalueserve to clean and harmonize their data. Evalueserve also developed a comprehensive data strategy to improve the logistics provider’s data management and return on investment (RoI). The robust data strategy transformed the client’s business, with over 90 percent adoption of the use cases Evalueserve identified.
The Challenge
Having already made significant strides in cleaning and harmonizing their data, the logistics provider knew it was time to scale their analytics efforts and move up the analytics maturity curve. The client asked Evalueserve to use AI to improve various elements of their business. To address their unique commercial operations challenges in the logistics space, they needed domain-specific algorithms rather than trying to force a generic AI package to work.
Our Solution
We began with trailer forecasting and customer segmentation because the two areas directly impact revenue and profitability for logistics providers. This helped their operations team make important decisions such as:
- How many trucks or trailers are needed at a particular distribution center today?
- How many drivers are needed at a specific distribution center today?
- How many people are needed at a specific distribution center today?
- What routes should trailers take?
Trailer forecasting helps companies increase profitability, reduce empty miles, and ensure trucks are close to or at capacity. Unexpected challenges can be mitigated with trailer forecasting. In addition, it allows companies to optimize supply chains with existing resources, which is critical as supply chains continue to face disruption from the pandemic.
Several types of AI techniques went into the trailer forecasting, including Time Series, ARIMA, XGBoost, Artificial Neural Networks, and Long Short-Term Memory.
Evalueserve performed customer segmentation of the logistics provider’s clients across all sectors. Customer segmentation helps businesses predict which customers are at risk of leaving (also called churn) and which may upsell to a client. The logistics provider now knows when they need to work to retain existing clients through offering incentives or proactively having more meetings with the client’s key stakeholders. Our customer segmentation efforts also tell the logistics provider which customers may be receptive to adding more product or service offerings to their current packages.
Evalueserve predicts LTVs for each segment, helping the logistics provider understand which clients stand to have the largest impact on their bottom line. To perform custom segmentation and predict churn and LTVs, our experts used algorithms including Time Series Forecasting, Logistic Regression, Cox Proportional Hazard Model, and Artificial Neural Networks.
Evalueserve also built out the client’s infrastructure to sustain AI and advanced analytics initiatives. We turned offline documents – including PDFs and physical papers – into usable data using optical character recognition (OCR) and chatbots. These alternative data sources allow the client to mine previously unstructured unusable data, adding valuable information for their analytics efforts.
We created a Machine Learning Operations (MLOps) team to continually and proactively solve any issues that may come up with the algorithms and maintain model accuracy. The existence of this team is particularly important, as algorithms can break or become less accurate over time, so being monitored for continual accuracy is a huge help.
Business Impact
Evalueserve provided trailer forecasting for 500 prioritized lanes (origin/destination combinations) and helped the logistics company improve its resource (trucks, drivers, people, etc.) allocation. Thanks to our work in that area, the client achieved a 3.8% year-over-year improvement in Linehaul On-Time delivery. That improvement led to increased customer satisfaction for the client and a ~$8-14 million impact on operating income.
Our efforts in customer segmentation, churn, and LTV resulted in a ~$1 million improvement in operating income and identified ~$52 million worth of revenue opportunity.
