Man + Machine = The Ultimate Product Discovery Team

By Troy Winskowicz, VP of product, Edgecase

These days it feels like “big data” is everyone’s new favorite bad word. E-commerce retailers are trying to manage big data to support large online product catalogs, monitor e-commerce performance, and craft awesome merchandising campaigns.

Not to mention, they have to get these things right in order for their e-commerce experience to be effective. At the same time, shoppers are becoming more accustomed to leveraging digital tools to find what they are looking for and answer the same questions they ask in-store. This leaves retailers struggling to bridge the gap between how customers found the perfect product before, (asking a human being) and the rising expectations of shoppers in the digital age (“Why doesn’t this website have ‘boyfriend jeans?!’”).

Case in point?  According to a recent survey by RichRelevance, 73% of consumers say they are extremely likely or likely to leave a retail site that doesn’t produce good search results. As well, 84% of e-commerce search engines don’t understand simple attributes like “cheap”, while another 60% don’t support thematic searches like “office chair,” according to the Baymard Institute.

So what’s a retailer to do?

The backbone of great merchandising, insightful analytics, and even simple site search is product data. Product data includes detailed product “tags” that correspond to a retailer’s product “categories.” Good product data can allow a shopper to search for a product that fits her specific needs, such as “daytime pink lipstick”, and find a few options to wear to work.  However, this type of descriptive long-tail search is only effective if the retailer has established a detailed library of product data for each product.

Developing enriched, robust product data at the scale, speed and quality that retailers demand is no easy feat. It requires expertise in handling daunting standard tasks and subjective problem solving. In other words, it requires a unique combination of man and machine, leveraging the super skills of both to deliver the product data needed to fuel robust e-commerce performance.

Machines At Work For Retail

Machines are experts at standard tasks. By incorporating machines into the role of the merchandiser, retailers can help create increased efficiencies and speed up the time-to-site for products online.  

For example, let’s consider a retailer who has a line of outdoor furniture that ranges from settees to lounge chairs. On her own, it may take weeks to classify each and every product into the appropriate category (chair, couch, settee, table, etc., etc., etc.). By leveraging a machine, the process of categorizing becomes much quicker. Auto-classification rules help rapidly assign each product to the correct category and technology can product accurate, efficient organization of data. This means more taxonomies and consistent search.

Humans, Doing What They Do Best

Humans are great at subjective problem solving. While they are not as efficient as machines at analyzing and organizing huge data sets, they are great at translating real world problems into workable ideas and using their creative and subjective skills to create enriched product data. Human merchandisers have unique insight about each of the products they sell, how shoppers are searching for them and what makes them special. They are also able to internalize the sense of excitement shoppers have when they find the product they are looking for.  They can then use that response as inspiration for creative campaigns and merchandising strategies that repeatedly bring customers back.

For example, a fashion retailer may be looking to promote their new line of dresses that are “Prom Perfect.” A knowledgeable merchandiser can look at each dress and decide if it’s “prom-worthy” or not, assign the proper attribution and add it to a collection. This might power a “Prom Perfect” marketing campaign featured on the homepage of the website, through email marketing, and social media promotion. While we eventually expect this type of attribution to fall to the responsibility of machines, today this type of creative, instinctual knowledge of dresses falls to the merchandiser, leveraging their skills for creativity and strategic thinking.

Working in combination, man and machine develop the rich, accurate product data required to get products online rapidly, and improve product discovery. This collaboration helps create and maintain the right product data that resonates with shoppers, helps achieve merchandising goals and builds credibility in the minds of shoppers.