Let's say you have a large inventory of coats - jackets, sweaters, parkas, anoraks, wind breakers and so on. You want a way to group them by their usage, but they all have different titles or brands - even though they serve the same function (protecting you from rain or keeping you warm).
They could possibly be buried in various parts of your catalog, so searching for “coats” may not necessarily lead you to a denim jacket, and searching for “jacket” may not give you a sweater. However, your customers may search for products by a particular name that is not associated with your products, and wouldn’t you want an easy way of showing them all options? Wouldn’t it be a shame if you stock the garment that fits customers’ needs but they can’t find it because it has a different title?
Our sophisticated models go beyond product titles, and look for patterns in categories and attributes such as size, material, functionality and usage.
With increasing complexity of product data, you can’t simply recommend another product for that particular product type, like you can with coats, because key attributes have to align in order for a product to be considered a true functional equivalent.
For example, if you are selling HVAC units and one requires 500 watts of power and another unit requires 5000 watts of power, this would not be a true equivalent.
Using fuzzy matching techniques and human intelligence, we handle various levels of complexity within product data.
The result
Receive consistent and accurate data regardless of the complexity and size of your catalog.
Our hybrid approach ensures consistent results, at scale, and is less expensive than the alternatives - that is, manual in-house effort or through an off-shore BPO. So even as you expand your range of products, you can keep the list of functional equivalents updated with no extra effort.
We put data first! Using our approach you can use data-driven insights to:
