Looking for B2B E-Commerce Leads? The E-Commerce Times Offers a 100% Satisfaction Guarantee Request Info




Posted By ECT News Network on 08/27/2019 in Artificial Intelligence

Retail Bundles, Recommendation Engines and AI: Fad or Future?

Retail Bundles, Recommendation Engines and AI: Fad or Future?

By Anthony Ng Monica 

Recommending the exact complementary product for each customer’s needs is a challenging task. It requires more than just an algorithm. It involves understanding the products being purchased, anticipating customers' desires, and expertly catering to what they want. This is something that online businesses -- big and small -- are trying to perfect, with retailers rushing to invest in new solutions in an attempt to find the answer.

Artificial intelligence, machine learning and other buzzword technologies are making the process more efficient and successful than ever before -- but are they enough to get perfect cross-sell bundle recommendations?

Is AI in retail a flash in the pan or a sign of things to come?

The Future: Better, Faster Recommendations

Retailers have understood the power of effective cross-sell for decades with their experiences in brick-and-mortar stores. One cannot blame retailers for their thinking process as they attempt to implement smarter tools. Better technology and more customer data should mean better recommendations – and they are better, compared to the traditional approaches of the past.

This is the function that sales assistants used to serve back when customers visited stores in person. Now this role is taken over by technologies that aim to do the same thing at scale and are calibrated to boost the retailer’s bottom line.

The new systems compare and contrast prior transactions and what customers previously appeared to consider purchasing. Armed with all this data, e-commerce retailers try to predict the most likely combination to result in an extra sale.

Artificial Intelligence improves the process by helping retailers to deliver recommendations at scale. However, it is not a perfect system, and there remain missed opportunities in the rapid rise of buzzword tech adoption for recommendation engines.

The Fad: Missed Opportunities

For more than a decade, retailers have relied on algorithms that cross-sell based on customer behavior. The fact remains that these recommendations often fail, however, and there is still plenty of room to improve. Retailers have been turning to hyped technologies, like machine learning and artificial intelligence, to fill the gap.

What they have been finding is that better technology does not necessarily solve the core problem of failed recommendations, lacking a true understanding of both the product being purchased and the customer’s needs. Additional data does help, but this remains an imperfect science. Amazon still does not have enough data in 57 percent of cases to make any recommendations at all.

Many businesses do not have the ability to collate enough data to compete with the likes of Amazon, so they turn to past customer purchases to inform their recommendations. This presents problems too, as it simply does not guarantee that the products will work together.

Just because one customer has purchased one item at the same time as another is not an automatic indication they are related. Customers can receive recommendation bundles for items that are not compatible, and therefore are unlikely to sell. Worse, the products may be returned by unhappy customers, hurting the bottom line.

The Outcome - for Now

Where does this leave us on the recommendation engine divide? Well, artificial intelligence and machine learning do show a lot of promise in this field. Recommendation engines are getting better, but there remain far too many cases of product pairings that do not make sense and do not convert.

Even with the best available technologies, recommendation engines continue to run into three main problems: irrelevant recommendations that look like they don’t make sense; missing recommendations on products; or incompatible products that cause unhappy customers to return products. Each of these scenarios hurts the retailer’s chances, and each will need to be improved before recommendation engines reach their full potential.

The recommendation engines of today fail to meet the most fundamental challenge when it comes to cross-sell. In order to deliver perfect cross-sell, it’s fundamental to take a step back to understand the products being sold. Retailers must ensure they’re compatible with the main product and relevant to the customer’s needs, and only then use artificial intelligence and machine learning algorithms to optimize the cross-sell offering.


About the Author

Anthony Ng Monica is founder and CEO of automated bundling service Swogo, which exceeded more than 1.5 billion cross-sell bundles in the first half of 2019.


Related Articles

Show Phone Number
818.461.9700

View Profile

Encino, California 91436

Search Blog Articles

Get the ALL EC Newsletter