Introduction
Drawing level with shrinkage and theft challenges through timely data and targeted strategies
Shrinkage and theft are pervasive and escalating challenges in the retail industry, with stores typically experiencing an average annual shrink rate of between 5% and 10%. Traditional monitoring methods often fall short, providing inaccurate early indicators and relying on semi-annual reporting that can quickly become outdated.
Retailers increasingly recognize the need for more timely and accurate data to inform loss prevention strategies across stores. It takes substantial time for retailers to swiftly act upon shrinkage issues as it requires tremendous effort to count every store item. Hence, a shift from long periods to short turns is critical for the effective allocation of resources and for implementing targeted strategies to curb this costly issue.
The challenge
Getting across the line with minimal losses
Our client, a leading off-price retailer, was grappling with significant losses due to shrinkage. They needed a timely and data-driven approach to resolve this challenge.
Specifically, they needed to shift from semi-annual indications to monthly decision-making regarding shortages, provide early alerts to stores and stakeholders, and accurately identify the top categories most susceptible to shrinkage across various stores and regions.
Footing an off-price retail dynamics advantage before the lights go out
Our sustained engagement has led to a deep understanding and extensive insights into internal operations, fostering expertise in off-price retail dynamics. As a result, we were well-equipped to develop tailored solutions for challenges encountered.
Solution
Engineering a hybrid solution that gave the client a pole position
Fractal engineered a hybrid solution that integrated Python code, Excel sheets, and Power BI reporting to automate the shrinkage prediction process.
This approach was chosen along with the client’s traditional method of physical counting, which often yielded misleading or outdated information. Our solution’s planning, design, and development phase took approximately three months and an additional two months for implementation and validation.
Beating shrink management through data-driven predictive modeling
The five-month timeline (three months to develop the solution and two months to implement and achieve the desired results) enabled us to provide a more timely, accurate, and data-driven approach to managing shrinkage through our solution.
We followed a four-step process designed to enhance prediction accuracy and operational efficiency.
| Steps | Actions | Output/Refinement |
| Data Gathering and Categorization | The team collected essential data, including scanning information, prior shrink records, and store attributes. Split data into three subsets based on shrinkage percentages for focused model training. |
Data preparation led to subsets for specialized model training, streamlining predictions based on shrinkage categories. |
| Data Cleaning, Model Training and Evaluation | After cleaning the data to remove nulls and other anomalies, a simple linear regression was trained on each subset, assessing model quality using metrics like p-values and directional shrinkage numbers as per their business implications. |
Model quality evaluations enhanced prediction accuracy and refined models for reliable shrinkage forecasts |
| Prediction Refinement and Adjustment | Calculated division-level correlations between actual and predicted shrinkage, making adjustments based on these values. For instance, if the division correlation exceeded 50%, the predicted shrinkage percentage was adjusted accordingly to create an alarm. |
Adjusted predictions resulted in more precise alarms and refined shrinkage estimates, thus fine-tuning predictions to make them more reliable and ensuring actionable and accurate insights. |
| Prediction Caps and Operational Implementation | Caps were also implemented- to eliminate negative or overly high seasonal predictions, resulting in more accurate and actionable final shrinkage predictions in both percentage and dollar terms. This helps in informing strategic decisions and resource allocation for efficient operational responses. |
Implementation of translated refined predictions led io actionable insights for better planning and resource allocation. |
Outcome
Immediate Impact
Our client swiftly reaped immediate and substantial benefits in their retail shrinkage management journey. They are periodically producing indications from our model while their conventional process helps verify and fine-tune our model actively. Our model with predictive accuracy soared to nearly 90% at the store level and 85% at the category level, unlocked enhanced inventory allocation strategies, and delivered invaluable insights, revolutionizing their operational game.
90%
Predictive accuracy at the store level85%
Predictive accuracy at the category levelIn the long term, the client will be able to pinpoint high-risk categories and stores, enabling better resource allocation, which, in turn, is expected to lead to a ~10-20% reduction in shrinkage within just one year.
Over time, the client gains the ability to identify high-risk categories and stores, optimizing resource allocation. This shift is anticipated to swiftly slash shrinkage by approximately 10-20% within a year.
Grab a PDF copy of this case study, making reference a breeze and sharing seamlessly convenient.
Download success story