What AI & Machine Learning can do for eCommerce?

This article was first posted on LinkedIn on November 4, 2019.

Several activities related to eCommerce and Digital Marketing can benefit from Machine Learning (ML) as well from other branches of Artificial Intelligence (AI). The integration of machine learning algorithms and eCommerce can help with all the main marketing goals: retention, engagement, and conversion.

Adoption of Machine Learning for Retail in 2018

There are many machine learning use cases in ecommerce, mostly limited to big players or startup with specific technical knowledge. Here is a list of the most significant areas of eCommerce that will eventually be enhanced or automated by ML:

  • Product Recommendations. This is the most typical use of ML for eCommerce. A recommendation system usually only requires the collections of orders and abandoned carts grouped by user. While proprietary recommendation systems are standard, we are starting to see Recommendation systems offered as a service for eCommerce sites. For example, Google Recommendations AI is offering recommendation services that can be integrated via API. If you don’t like to share your data with Google, take a look at Recombee.
  • Marketing eMails. ML can help create personalized messages and optimize frequency and CTA for each customer and email message. ML can deliver results much faster and more effectively than simple A/B testing. For example, by using ML to tune their email messages, Dell saw a 50% increase in email CTR, while Harley Davidson saw a 40% sales increase.
  • Notifications Management. Similarly to the mailers, ML can define frequency, time of the day, content, CTA, and content for notifications. Working with ML, in 2018, Facebook was able to recover most of the recent drop in DAU (Daily Active Users) by using ML to select the frequency of their app’s notifications. One startup working on notifications optimization is Gradient.
  • Personalized Navigation and User Experience. The website and the app will adapt to each user by offering options and selections that are consistent with his/her history, preferences, and personality. The main goal is to personalize the website consistently with the user profile and intent. According to a Accenture report, 75% of consumers are more likely to buy from a retailer that recognizes them by name, recommends options based on past purchases, and knows their purchase history.
  • Ad Campaign Management. It is already happening. On Google Ads, we have the option to let an ML system in Google Ads to manage a campaign optimizing the set of keywords and the bidding process. Better and more relevant ads, benefit the advertisers, and also the search engine.
  • Copy and Content Generation. We see the first wave of startups, like Phrasee, working on automatically creating content for marketing emails using a subfield of Artificial Intelligence called Natural Language Generation (NLG). Another product in this market is Ginnie, a simple module for Shopify, that automates the copy generation. At the moment, it’s minimal, it doesn’t apply any knowledge specific to the store or customers. Content generation is used to improve conversions and SEO performance, helping sellers generate unique content that appeals to customers as well to Google. A few more exciting companies to watch in this area: Narrative ScienceMash’n LearnCrew Machine, and Arria.
  • ChatBot Automation. It’s an exciting application of ML and Natural Language Processing (NLP) that can touch different areas, including sales, tech support, and customer service. There are too many companies in the US to name them. Search on google for “chatbot”, and you get the whole list. If you are in Italy, I would recommend taking a look at b-Optimist.
  • Shipping and Inventory Management. A Machine Learning system can assign each order to a specific warehouse based on historical data, product stock, and user delivery preferences. Even the Inventory can be managed with ML to better predict the timing of purchase for every product while minimizing the cost of the excessive stock. The entire Supply Chain system is impacted by ML.

At the moment, some of the areas of impact for ML listed above seem to be unapproachable by small sellers, but I expect this to change quickly. In the meantime, new eCommerce companies are already organizing their entire eCommerce business around the Machine Learning technology. Among the most well-known brands, StitchFix is an excellent example of extensive use of ML and algorithms.

More eCommerce Articles

Franco Folini lives and works in the eCommerce territory, a wild area between the Kingdom of Technology and the Kingdom of Marketing. He speaks fluently the language of both realms. For many years, Franco has been helping people from both sides to reach across the aisle and successfully collaborate.

If you want to find out more about Franco, visit his LinkedIn profile or send him an email folini[at]gmail.com