Online retailers struggle to capture shoppers’ attention. After all, it doesn’t matter if you have a great product in stock, if the right shopper never sees it.

In stores, merchandisers use their market knowledge to highlight the most attractive products. Sales associates add a personal touch; they learn an individual’s needs and retrieve the best products for her. This model has been successful for generations.

Shops Sign

Online shoppers demand the same sort of personalized attention in the digital store. Dazzled by Facebook and Netflix, they expect websites to interpret any signals they’ve shared and serve a relevant experience. This has put a considerable burden on digital merchandisers — how can they show this personalized care to the thousands of shoppers visiting their online stores? U.S. retailers look to Silicon Valley for the cutting-edge technology that can address this problem while redefining the role of the online merchandiser in the process.

Online merchandisers must force innovation

Algorithms are great optimizers. An algorithm can detect patterns in unstructured data and serve an experience that is relevant and personalized to the user. This enables companies to feed their “big data” into such algorithms and jump-start their personalization efforts.

But as impressive as they are, machines and algorithms don’t innovate. That’s where humans come in. They bring the art to merchandising.

Merchandisers understand how shoppers will react to the latest market trends. They see upcoming product releases and assess how they will change shopping patterns. Whether they’re using a data-driven process or relying on experience, this ability to predict behavior is the merchandiser’s key strength. Of course, once they predict consumers’ behavior, the next step is to act based on these insights.

Traditionally, a digital merchandiser would try to create a one-size-fits-all campaign online for a new product release. There were trade-offs to be made. What if the new product was from a brand a particular shopper couldn’t afford? Should the merchandiser still take prime real estate on the page for this product? The merchandiser would weigh the decision and if the answer was yes, she would publish a particular product assortment that would then be shown to every visitor on the site.

With data-driven personalized search, however, a merchandiser can leverage algorithms to her advantage. An algorithm can easily detect a shopper who never buys men jeans. Based on this, female jeans may be emphasized for this particular shopper. Each action a user takes on a page informs the algorithm of what the shopper is willing to buy. The algorithm uses that information and the merchandiser’s work to optimize the end user’s experience. This combination of human and machine maximizes revenue for the retailer and improves the user’s experience. 

Savvy merchandisers identify gaps in algorithms

Every algorithm has its limitations. In particular, algorithms often fail to optimize experiences in the scenarios listed below.

  1. Insufficient data: Consider a search algorithm that uses online sales data to more prominently show products often purchased after a shopper uses a particular search term. The algorithm will be able to optimize quickly for popular search terms. However, for search terms rarely seen, or new trends, it could take longer for the algorithm to properly optimize the experience.
    For example, a merchandiser notices that “snow jackets” is not a popular search term. But the results for “snow jackets” optimized by an algorithm aren’t great. The merchandiser knows that there are big snow storms coming. They look at aggregate data for similar terms such as “winter jackets” and “coats”.  Using this data, they boost popular products for the “snow jackets” searches.
  2. Redefining events: Disney released the movie “Frozen” in November 2013. The movie was a huge success, completely redefining the term “frozen”. An algorithm using historical data might react to the term “Frozen” by promoting frozen food, for instance. But frozen food is no longer relevant to consumers who are looking for items related to the hit movie. Merchandisers act on this immediately. A merchandiser can add a rule to boost Disney products across the site for searches including “Frozen.” Algorithms need input by a merchandiser, or enough volume of big data, to determine that these two phrases are no longer related.
  3. Temporary business considerations: Traditional goals for e-commerce sites are maximizing revenue and profit. There are times, however, in which these goals may temporarily shift. Take the example of a fashion retailer who is about to launch their spring catalog. A merchandiser can influence the site’s ranking algorithms ahead of the launch to promote the current season’s products with high inventory. This merchandiser’s insight can help reduce the need to place products on clearance after the launch.  

As retailers capture new information about their shoppers, data-driven merchandisers can devise new ways in which to use data to complement current algorithms. Through these efforts, they are able to fundamentally improve the core algorithms responsible for the experience of tens of thousands daily shoppers. Merchandisers who successfully leverage algorithms can push the boundaries of what can be accomplished in online retail — and they can guarantee that when they have a great product in stock, the right shopper is going to see it.

Shop sign photo by Steve Snodgrass published under Creative Commons license.

Omar Fernandez is a BloomReach product manager.