How Does Centricity’s Platform Function?
What differentiates Centricity from competitors is its ability to provide insights on pre-purchase consumer intent. The company’s focus is to help brick-and-mortar retailers to optimize the assortment of their products down to the stock-keeping unit level. In fact, it is necessary to create an assortment that matches demand, because digital native retailers such as Amazon are already doing it in real time. On the other hand, it’s focus for manufacturers is to help them maintain their relevance in their market by having the right tools at their disposal.
Centricity tracks web page views by consumers across e-commerce, news, recipes, blogs and web search to find out what people are searching for. In one day, the company claims to record over 1 billion-page views—which translates to location, time, and URL data. In the next step, the company leverages both human intelligence and machine learning to identify the URLs that receive the highest traffic and label them with products and consumers’ intent to purchase. Centricity is thus able to understand real-time consumer purchase intent for a wide range of products.
In addition, the company applies correlation analysis against sales data to identify the correlation between sales data and consumer intent. Figure 6 explains the functioning of Centricity’s platform.
In order to label products and consumer intent, Centricity leverages a human-in-the-loop, natural-language processing system that operates in the following three distinct phases:
- Entity Extraction: In this phase, Centricity identifies relevant products on the page. For example, chocolate is an entity in the sentence: “Eating chocolates increases the risk of diabetes.”
- Sentiment Analysis: For each instance of an entity being identified in the first phase, Centricity determines the sentiment on a scale of -5 to +5, which corresponds to consumer intent to purchase a product or not.
- Classification: With all the entities, in parallel to the sentiment analysis, Centricity performs standard classification and assigns each instance to an entity in its hierarchy. Once the classification is done, Centricity’s platform assigns an intent score to each entity by averaging the sum of scores of all instances. For example, if there are three references to “chocolate,” a +1, a -1, and a +3, the algorithm would generate a score of “+1” for every view of that page.
Centricity: Competitive Advantage Matrix
In this section, we analyze Centricity’s market positioning, current opportunities in the market and growth drivers, and the company’s unique attributes.
First, we identify some of Centricity’s key competitive advantages below:
- Accuracy: Centricity leverages big data architecture to track over 1 billion intent events each day, while geolocating each one of them to within a one- mile radius—half-mile in a few cases—making the data highly accurate when identifying consumer intent across regions.
- Pre-purchase shopper behavior: Centricity helps retailers and manufacturers understand pre-purchase consumer behavior, which allows them to optimize their supply chain operations.
- Data Unification: Centricity’s platform leverages data from multiple sources to create a picture of future shopper demand based on their interest across different platforms.
- Flexibility: Centricity’s platform can be configured for individual businesses.
- Real-time insights: Centricity helps its clients by capturing consumer purchase intent in real time, enabling them to plan for relevant substitutes before the demand of a certain product plateaus.