Embedded matching for a perishable-goods platform
How a vertical platform automated product matching across 100+ live supplier feeds
A vertical software platform serving distributors of perishable and farm-sourced goods needed to reconcile a constantly changing inventory with product feeds from more than 100 upstream farms and suppliers.
The platform handles more than 500,000 live SKU records. Within that catalog, approximately 300,000 records must be continuously matched between distributor inventory and changing supplier feeds.
Product identifiers were frequently missing, product names varied between suppliers, and employees still resolved uncertain matches through phone calls and spreadsheets. The process was slow, difficult to scale, and had to be repeated whenever supplier data changed.
Claro was embedded into the platform as a continuous product-matching and product entity resolution layer.
At a glance
Industry: Perishable-goods and fresh-produce distribution
Operating model: Vertical software platform serving distributors
Supplier network: More than 100 upstream farms and suppliers
Catalog scale: More than 500,000 live SKUs
Matching scope: Approximately 300,000 SKUs matched continuously
Geography: Global sourcing with European operations
Measured result: Approximately 40 hours of manual work saved per month
The challenge: matching a catalog that changes every day
In perishable-goods distribution, a catalog never remains static for long.
Availability changes daily. Suppliers may use different names for the same product. Units and descriptions may vary. Stable product identifiers may be incomplete, inconsistent, or entirely absent.
This creates a difficult product-matching problem: determining whether an incoming supplier record refers to an existing product in the distributor's inventory or represents a genuinely new product.
Before Claro, much of this reconciliation still happened manually. Teams compared spreadsheets, contacted suppliers, and resolved uncertain records over the phone.
Every new supplier feed created additional manual work, and previous matching decisions were difficult to reuse consistently.
Why exact identifiers and static rules were not enough
Traditional catalog matching often depends on exact identifiers such as GTINs, supplier SKUs, manufacturer part numbers, or standardized product names.
In this environment, those identifiers were frequently weak or missing.
Exact matching alone could therefore miss valid relationships, while broad text similarity could create false matches between products with similar descriptions.
The platform needed a matching process that could:
Combine multiple product signals instead of relying on one identifier
Work with incomplete and inconsistent supplier data
Assign a confidence score to every proposed match
Preserve the source and reasoning behind each decision
Route uncertain cases to human review
Re-run automatically when supplier feeds changed
Write approved results back into the existing platform
The solution: embedded product entity resolution
Rather than building and maintaining a complete matching engine internally, the platform embedded Claro into its existing data workflow.
Claro operates between incoming supplier feeds and the platform's canonical product records.
When a supplier adds or changes a product, Claro identifies the affected record, standardizes the available data, generates potential matching candidates, and evaluates whether the supplier record refers to an existing product.
High-confidence matches can continue through the workflow automatically. Uncertain records are sent for review before any result is written back.
This gives the platform a reusable matching layer without requiring its team to build product entity resolution infrastructure from scratch.
How the matching workflow works
Detect new and changed supplier records
Claro identifies additions and updates as supplier feeds arrive, allowing the matching process to focus on the records that require attention.
Normalize inconsistent product information
Available names, identifiers, units, descriptions, and supplier-specific fields are prepared so that records can be compared consistently.
Generate candidate products
Each incoming supplier record is compared against relevant canonical products in the distributor's existing inventory.
Resolve product identity
Claro evaluates multiple signals to determine whether the incoming supplier record and an existing distributor record refer to the same real-world product.
Validate confidence and provenance
Every proposed match includes a confidence assessment and the evidence supporting the decision.
Low-confidence or ambiguous cases are routed to a review workflow instead of being accepted automatically.
Write approved records back into the platform
Trusted matching results flow back into the platform's product data model.
The process runs again as supplier feeds change, keeping the matched catalog current rather than treating reconciliation as a one-off data-cleaning project.
Results
The platform replaced a substantial part of its phone-and-spreadsheet reconciliation process with a continuous embedded matching workflow.
The implementation now supports:
Approximately 300,000 SKUs matched continuously
Product reconciliation across more than 100 supplier feeds
Approximately 40 hours of manual work saved per month
Faster integration of new supplier records into the matched catalog
More consistent canonical product records
Confidence-based review for uncertain matches
Product matching delivered as a native platform capability
Continuous re-evaluation as supplier data changes
The platform can offer matching directly to its distributor customers without having to build and maintain its own entity resolution system.
Why continuous matching matters
For multi-supplier distributors, the challenge is not simply storing product data.
The harder problem is maintaining a trustworthy product identity while suppliers, availability, descriptions, and identifiers continuously change.
A one-time catalog cleanup becomes outdated as soon as the next supplier feed arrives.
Continuous product matching allows the platform to preserve a reliable canonical catalog even while the underlying source data changes every day.
Key takeaway
The platform did not need another static product database.
It needed an operational layer capable of continuously resolving product identity between distributor inventory and changing supplier feeds.
By embedding Claro, it transformed product matching from a recurring manual task into a scalable platform capability.
Frequently asked questions
What is product matching in a multi-supplier catalog?
Product matching determines whether records from different suppliers and internal systems refer to the same real-world product, even when their names, units, descriptions, or identifiers differ.
How can products be matched when identifiers are missing?
Claro combines the available product signals, generates potential candidates, evaluates each candidate, and assigns a confidence level. Uncertain matches can be reviewed by a person before being written back.
What is product entity resolution?
Product entity resolution is the process of identifying which records across different datasets represent the same product and connecting them to one trusted canonical record.
Can product matching run continuously?
Yes. Claro can detect new and changed records, re-run the matching process, and update approved results as supplier feeds evolve.
Can product matching be embedded into another software platform?
Yes. Claro can operate on top of an existing platform's data model so matching can be delivered as part of the platform's own product experience.
Does every product match need manual review?
No. High-confidence cases can be processed automatically, while ambiguous or low-confidence cases are routed to review according to the customer's rules.
