How Unilever Uses AI to Make Winning Products
Unilever is among the world’s largest consumer goods companies. For nearly a century, Unilever has been steadily growing and expanding around the world, now with an annual revenue of over $60 billion.
Today, Unilever products—which include brands in foods & refreshment, beauty & personal care, and home care—are used by over 2.5 billion people. At Unilever’s scale, it’s possible that billions of dollars have been spent on R&D for product innovation.
Unilever’s product teams are tasked with creating high-quality products that are loved by consumers, across a number of well-known brands, which is an incredibly difficult task when faced with such a massive amount of products and consumers.
Better Product Strategy
Traditional product strategy is a difficult task, and traditional means can’t adequately analyze the variety of platforms customers use to create a holistic understanding of sentiment, which drives sales.
Therefore, Unilever came to Commerce.AI with three goals in mind:
- Analyze leading products & brands
- Interpret reasons for success
- Gain recommendations for new product strategies
Commerce.AI’s data engine was used to scan and analyze nearly half a million reviews on Amazon US, across over 130,000 products, in order to achieve those goals. The Commerce.AI data engine automatically collects, cleans, reads, and analyzes text data in order for brands to leverage sentiment for product strategy.
The insights from this analysis were merged holistically with previous analyses on over 600 million products and services from over 100 sources.
First off, the data engine was used for category analysis, which means automatically detecting leaders in any given product category. The algorithms then generate attribute trends for the overall category - with an attribute being an aspect of the product or market that drives a consumer’s purchase decision in the given product category.
Additionally, the data engine was used to analyze sentiment market share, and find the top-performing brands according to the number of reviews and their share of positive sentiment. After sentiment market share was discovered, product-related pages were used to find the reasons causing positive feedback. Finally, human judgment was used to create a set of high-level recommendations.
Results
Based on the analysis, Commerce.AI identified gaps in the portfolio of Dove (a Unilever brand).
These insights helped to inform new product launches and acquisition strategies.
The Commerce.AI platform empowers the decision-makers with large-scale structured and quantified data derived from unstructured feedback provided by real customers, resulting in accelerated AI-driven product decisions.
How it Works
We’ve explored how Unilever uses AI to make winning products at a high-level, but how exactly does it work?
Commerce.AI automatically scans the web – including customer reviews and feedback – to identify new and trending products and category attributes, across own and competing brands, that can be used to deliver a more relevant and satisfying customer shopping experience. These insights are built on a framework that supports multiple data formats to provide a robust understanding of the product quickly and efficiently.
Commerce.AI’s solution gathers public opinions about products in forums, on YouTube, or in articles. The results are visualized in a self-service dashboard. Based on that reports can be created that show insights about strengths, weaknesses, opportunities, trends of specific product segments. Typically, a quick feasibility study is done to ensure that Commerce.AI's data engine can be used to improve a company's product strategy. During that phase, Commerce.AI also investigates if there is enough data to feed their artificial intelligence models.
Summary
Product teams today have a harder job than ever before, and are tasked with creating successful products in a highly-competitive market.
Customers have more choices than ever before, but are also generating more data than ever, which means that AI-powered product teams can conduct better market research to understand customers, and stay ahead of the curve.
