Why Retail Giants Still Get Product Recommendations Wrong
Bad recommendations hurt e-commerce. Online retailers must be armed with product and customer data to correctly recommend complementary products. If they don't have the data points, nor the ability to use them, mismatched product recommendations are inevitable.
This can be a source of embarrassment for even the biggest retailers. Amazon.com made the headlines recently for all the wrong reasons by bundling a school backpack and a kitchen knife. The fact that one of the world’s biggest retailers struggles to recommend relevant and compatible bundles like this is more than embarrassing, it's a source of lost sales.
Retail giants are missing a trick by relying exclusively on customer behavior and simple rule-based systems. Let’s explore what retailers can do today to ensure bundle success tomorrow.
There's great opportunity and great challenge in today’s retail landscape. Gross margin has been squeezed in what's the most promotional market in almost a decade. The rise of e-commerce offers more sales opportunity, yet conversely translates into shoppers who expect year-round deals and competitive prices.
The easiest way to increase gross margin is to simply sell more to the customers you already have — and retail giants consistently turn to product bundles to make this reality. However, many retailers continue to use flawed product bundling systems. First, there are manual suggestions which do not scale. Second, there are recommendation engines which act on customer behavior. This system, however, fails to understand the products, and therefore deliver random recommendations.
As demonstrated with the knife and backpack combination, products “frequently bought together” do not guarantee compatibility as they rely on customer behavior. Simply because two products are purchased at the same time doesn't mean the next customer wants the same. Perhaps with the school backpack, it’s more likely a customer will want a pencil case or stationery.
Irrelevant recommendations cause unhappy customers, product returns and, perhaps worst of all, don't sell. Interestingly enough, Amazon still doesn't have sufficient sales data in more than half of all cases to offer product bundles.
Why Product Recommendations Matter
Retailers need to fine-tune their recommendation engines if they're to recoup those lost opportunities. In this way, understanding the product needs to be at the core to guarantee compatible pairings. A smartphone with a case or battery addition not only makes sense but also presents the chance to pair with high profit margin products.
Technologies such as artificial intelligence and machine learning are becoming popular tools in the bundling industry. However, smart systems make no difference if the retailer fails to understand its products. Bundling systems need to learn from what does and doesn't work, and paired products need to drive relevance.
Smarter Systems, Higher Values
Adding related products to a pending purchase is a powerful way to grow revenue. One estimate from McKinsey suggests that 35 percent of all Amazon purchases come from recommendations. Imagine that figure if the system improved suggestions and increased success.
There are a few ways retailers can do this. One method is to offer free shipping to encourage customers to build custom bundles and spend more, while also allowing companies to carry low-cost items profitably.
Then, of course, there's data. More information makes for smarter systems — something which bundling must achieve if its potential is to be realized. Retailers need to understand their products; that's what they're all missing. They all have customer data, but this doesn't guarantee the recommendations make sense.
One should expect the trend towards data-driven recommendations and decisions to continue, and this means investment in bigger and better recommendation engines.
Anthony Ng Monica is the CEO of Swogo, the world’s first automated bundle solution for e-commerce retailers to increase margin. Hundreds of retail leaders in over 30 countries around the world drive profitable growth with Swogo. Swogo takes a unique approach that focuses on understanding a retailer’s product assortment - Swogo Product Graph combined with machine learning and AI algorithms surpassing billions of recommendations per year.