Efficient MRO Data Classification & Cleansing for Shipping Operations - Achieving 100% Accuracy in Categorizing Over 196,000 Items

MRO Data Cleansing & Enrichment

Prior to classification, the product data was cleansed and enriched to ensure its quality and accuracy. This involved identifying and correcting discrepancies (e.g., missing part numbers, incorrect descriptions), enriching the data with the latest product attributes, and standardizing it to ensure consistency across all entries. This process ensured that the final classification was based on clean, accurate, and standardized data.

Item Description Analysis

The next step in the classification process was a thorough review of the provided item descriptions. This step was crucial in understanding the product type and its corresponding classification. Each description was carefully analyzed to ensure that the item was accurately classified based on its usage, characteristics, and specifications in the MRO context. This allowed the team to assign appropriate categories, reducing errors in classification and increasing data reliability.

UNSPSC Classification

We assigned a unique UNSPSC (United Nations Standard Products and Services Code) to each item based on its product type and function. The classification process was driven by an in-depth understanding of the shipping sector, including mechanical parts, ship repair equipment, navigation tools, and accessories. Our team ensured that each MRO part was mapped to the appropriate UNSPSC category, aligning with industry standards for seamless integration across inventory and procurement systems. This classification provided a structured, standardized approach to managing the large volume of items, facilitating easy retrieval.

Assigning the Correct Unit of Measure (UOM)

When assigning UNSPSC codes, we ensured that the appropriate Unit of Measure (UOM) was considered for each item. This step was essential for accurate classification, as the UOM directly impacted how the product was categorized within the UNSPSC system. For example, items like mechanical parts, fasteners, and seals were classified with a UOM of "pieces," while larger equipment, such as ship engines, propellers, or navigation systems, were assigned a UOM of "units" or "sets." This careful attention to UOM ensured that each item was correctly mapped to its respective category.

Commodity-Level Classification

The project required strict adherence to commodity-level classification, which meant that each item had to be categorized in precise detail according to its specific function and characteristics within the shipping industry. For example, instead of broadly classifying items as "mechanical parts" or "marine equipment," each part was categorized into precise subcategories like "engine components," "hull fittings," "mooring equipment," or "marine pumps." This granular level of classification ensured that no items were misclassified or grouped under generic categories, facilitating accurate inventory management and enhancing the client's ability to quickly locate and manage MRO parts and other critical shipping components.

Quality Control (QC) Process

To ensure data accuracy, a multi-level quality control process was implemented throughout the classification effort. The process involved several stages of checks, with each level playing a crucial role in maintaining data integrity:

• Level 1 - Self-Verification: The initial quality check was carried out by the doer themselves. The individual responsible for classifying each item reviewed their work to ensure that the classification was accurate and met the required standards.
• Level 2 - Group Head/ATL Review: Once the item classification was completed, the Group Head or Assistant Team Leader (ATL) conducted a second-level review. This ensured that any inconsistencies or errors were addressed early in the process.
• Level 3 - Account Manager Random Quality Checks: The final level of quality control involved random spot checks by the Account Manager to verify the overall accuracy and consistency of the classification process, ensuring that the data met both internal standards and client expectations.