We provide data science services to support decisions and help human judgment in developing business strategies. Through predictive models, we evaluate the effects of real-world variables in order to predict a complex phenomenon. Through prescriptive models, we suggest the best course of action following criteria established by the previous analysis.
Solutions developed for our customers encompass several aspects of customer experience and customer intelligence: automatic segmentation of the (current or prospect) customer base, churn, up-selling and cross-selling models, the resultant customer ranking. We also study the latent behavior of the population on the basis of the big data we can measure (e. g. online searches).
We develop: sales forecasters by integrating historical data with other significant variables and recommender system for products, social media profiles (for example suggesting new contacts and influencers), store assortments. We also work on cluster analysis for the customer intelligence, automatic anomaly detection (e. g. retail best sellers, cyber security attacks), identification of trending topics and other emerging phenomena in social networks.
We are currently focusing our Data Science activities on two main fields: price intelligence and retail supply chain.
One of the most important Italian insurance group relied on CELI for the segmentation of the customer base. Through descriptive data analysis, CELI can characterize customers and profile them using clustering algorithms. This information allows us to understand customer behavior and define personas, each with his own priorities, in order to help the company better satisfy the needs of all their customers.
A leader credit company asked CELI for help in improving the definition of their pricing strategy. What we do is support the operators in market monitoring, giving them a statistical support to determine their pricing. Thanks to our technologies, we can elaborate an automatic pricing plan based on both the human know-how and the knowledge inferred through machine learning. We provide operators with an interface which allows to monitor prices and compare their pricing plan to that of their competitors, and provides forecasts on the number of sold contracts.
A world-famous Italian luxury fashion brand turned to us for help in making their retail supply chain operations more efficient, by improving the accuracy of their sales forecasting, better identifying unexpected best sellers and slow movers (products that sell less than attended), and optimizing product supply to each store.
We integrated internal sales, product and customer data with external data sources about location intelligence, social media content and product description on web catalogues and verified the predictive potential of all these data sources.
We then used this data to train predictive models able to provide sales forecasts, which are then monitored through an interactive dashboard with automated alerts, and suggest an efficient product assortment for stores.