Taiwan's e-commerce platforms have enjoyed leaping revenue growth since the COVID-19 outbreak as consumers increasingly shop online instead of going to brick-and-mortar stores. According to the Ministry of Economic Affairs, online retail sales reached NT$503.5 billion in 2023, of which NT$284.2 billion, or 56% was generated by pure-play e-commerce companies. To optimize the shopper experience, e-commerce operators have begun to use AI to analyze customers' purchase history so as to make recommendations when they browse the shopping site thereby increasing the likelihood of a transaction.
According to Rosetta.ai CEO Daniel Huang, e-commerce operators connect customers with products using tracking and recommender systems to predict buyer behavior. These systems growingly adopt AI to more accurately capture consumer preferences as AI technologies continue to advance. Rosetta.ai's solution is unique in that it matches the characteristics of both people and products. On top of that, it has the capability to understand images and texts, making it especially suitable for fashion e-commerce platforms. It is being used by multiple e-commerce companies.
More than 3,000 use cases with no need for a single line of code
Founded in 2016, Rosetta.ai started out with a focus on developing tracking and recommender systems for e-commerce platforms. As it builds up its capability in AI and machine learning, Rosetta.ai has created a unique merchandise tagging database for fashion e-commerce companies that can be used throughout every stage of the shopping journey to realize automated marketing, including precisely redirecting target traffic, enabling diverse personalized interactions and increasing the repurchase rate.
In line with the no code or low code trend, the use of Rosetta.ai's recommender system requires no coding. With only simple drag-and-drop operations, users can have personalized recommendations displayed anywhere on the shopping site, whether it's the home page, a merchandise page, a category page, or the shopping cart page. Seven recommendation scenarios are available to significantly increase the hit rate. Since its establishment, Rosetta.ai has helped more than 3,000 e-commerce brands grow their revenue.
Huang mentioned that fashion e-commerce companies often use a large quantity of images to show how well clothes and accessories go together. Cosmetics retailers do the same to demonstrate how good their products would make consumers look. Traditional recommender systems are generally not capable of performing in-depth analysis and often tend to over-recommend similar products. For example, the system will keep recommending black clothing to a shopper who has purchased a black item. Rosetta.ai, on the other hand, recommends items that go with what the shopper has bought based on his or her purchase behavior. For example, it will recommend pants, skirts, handbags, or hats that go with the black top the shopper has bought. Its unique deep learning-based image analysis has attracted many fashion e-commerce companies.
Using AI-powered precision recommendation algorithms, Rosetta Engage gains insights into every shopping journey touchpoint and offers intelligent recommendations through three interactive experiences. By accurately identifying moments of interaction between shoppers and website products, it is able to precisely recommend items shoppers love and want to buy, thereby raising the conversion rate.
Expanding into Japan, Rosetta.ai's globalization efforts are generating impressive results
Looking back at his entrepreneurship journey, Huang said that Taiwan's startup environment was not yet mature when he founded Rosetta.ai. There were no accelerators to offer guidance and limited sources of funding. Luckily, Rosetta.ai was subsidized by the Ministry of Economic Affairs' Small Business Innovation Research (SBIR) program. Taiwan Tech Arena (TTA) also helped Rosetta.ai plan marketing strategies and secure funding. Rosetta.ai now has a team working in Japan to strengthen collaborations with local fashion e-commerce companies. It also plans to incorporate a Japan-based subsidiary sometime between late 2024 and early 2025 as part of its all-out efforts to foray into Japan.
Huang noted that apart from the recommender system, Rosetta.ai has launched Rosetta Traffic, a whole new ad model that leverages AI to create a personalized shopping environment and drive high-spending consumers to the shopping site. It also enables product exposure on complementary brands' websites to increase traffic and improve the average order value. Rosetta.ai expects Rosetta Traffic to become its growth engine in the future.
Shopping journey analysis increases traffic and revenue for small- and medium-sized businesses
Photo: Rosetta.ai