Product data is available in data feeds and on web sites. In order to create Quality Data bad data must be detected in data feeds and on web sites. The bad/missing product data must be recognized in order to make good matched records.
I found searching for products on the web infuriating because:
- Sites failed to find the products I search for.
- Sites provided incorrect and incomplete information about the products they sold.
- Related products were hard to explore, e.g. different trim/color/sizes etc.
The solution to many of these user experience problems are matched product records. A matched product record is the same product found at different web sites, including variants.
Manufacturers create product records. Retailers, agencies, and aggregators transform the records. The transformation process introduces errors, removes data fields, and moves data from one field to another.
The same record at different stores can contain different model numbers, UPC, product images, … Detecting the bad/missing data in the records is necessary in order to make matched product records.
Bad data at one store can lead to all sorts of problems during the product matching stage. Below is an example of products which match because all products contain the same UPC at one store. The bad UPC must be identified and removed in order to match the records correctly.
Internet businesses are leaving lots of money on the table if they do not fix these problems. The key to addressing these problems is detecting bad data and matching good data, including variants. The result is accurate, well-structured product data.
Data Record Science constructs an up-to-date high-quality Universal Product Database by automatically combining and correcting product information from data feeds and retailer sites. Data Record Science’s mission is to provide quality data using matched product records which will increase revenues for Internet companies.