There are many parts of the online shopping process and all have to work well in order for the shopper to have a great user experience. It only takes one bad spot in the long chain from the first contact with the shopper to supporting the customer after the purchase to lose the customer, but also the bad reviews and social sharing of that customer that result.


One area that the retailer often overlooks is the role of product catalog in the customer experience. Perhaps this is because the catalog is often thought of as an IT issue, whereas marketing/advertising and user experience are seen as being the responsibility of marketing. Savvy retailers, though, are realizing that every interaction with the customer AND the data and processes that support those interactions are critical.

How does the product catalog affect the user experience?

Our experience in working with retailer online catalogs is that many have both a small number of product attributes defined in the database data dictionary and, for those that are defined, many of the structured data fields are left empty. For example, the retailer may have a catalog database field for color, but that field is not filled in for many items. Why is that important? It is important since those database fields are used as part of the guided navigation and search processes. Rather than being able to get the items from a database query of the structured fields, the only choice is to do keyword searches of the prose in the product name and product description. The end result, however, is that the shopper may not find the items even though the store does have items that do match.

Let’s look at the example below. Using BestBuy’s[1] mobile app or mobile website, I type in “red laptop” and I scroll results “1-10 of 140” to see 10 laptop bags and backpacks. When a search does not find relevant items on the first page, I have been “well trained by Google” to change the search word string to try to be more relevant. So, I try “laptop in red”, but get the exact same results. At this point, most shoppers would assume that BestBuy just does not have any red laptops. Few would retrieve the next page to see results “11-20 of 140” and see 2 red laptops and an additional 5 laptops in mostly black, but with a bit of red.[2]


Is the problem just that the information does not exist? Often that is not the case, that is, the information is in the product database, but in places such as the product name or product description. However, it has not been entered in structured data fields where the standard search and guided navigation can make best use of it. The keyword search of the product name and description finds the word, but it is difficult for the search algorithm to judge the relevance to the search. The underlying reason is often that the information is imported from product data from the manufacturer or distributor.

How Vioby has solved this problem

How can this problem be solved? The obvious solution is for the retailer to have people go over the catalog data for each product and fill in the data fields appropriately, but this can be a very expensive option for all but the most popular and best selling items.

We believed that the human intelligence that a person would use to improve catalog data manually can be integrated into automated algorithms based on machine learning and natural language processing that would accelerate and simplify the process. Our knowledge-based solution better classifies the product attributes from all the data in the catalog and other sources, so that when a shopper is looking for products that meet a set of requirements, the shopper can find those items easily. This results in improving another interaction that is part of creating a great user experience for your shopping customers.

Want to know more? Just email me. Or you can click here to request a demo of our voice-interactive shopping assistant.


[1] BestBuy’s search engine is not at all unique with this problem. It was used only as an example that seemed clear and simple.

[2] If I were on their website for desktop and laptops, the default is to view 15 per page (rather than 10 per page on the mobile. Then, if I scrolled to the last 5 items on the page, the visible items would now include 1 red laptop and 2 laptops in black with some red trim.