Challenge

Help a Large fashion retailer that was taking 36hrs to process 2 years of ePoS data (8 Billion records) within the traditional business intelligence system. (Microstrategy & Oracle). Merchandising team required predictive analytics and better performance to enable calculations on sales data to enable stock to be in the correct stores for the following week. (Calculations would include last week’s sales this year compared to next weeks sales last year).

Solution

fraXses deployed an in-memory columnar database on top of the fraPses platform into which all the ePoS data was fed on a real time basis. All the analysis was also completed in real-time with results being delivered in less than one second.

The system was then extended to incorporate inventory and weather data, based on store locations, to provide a more accurate analysis of factors driving sales.

Result

The retailer has been able to respond to localised demand for inventory in a fraction of the time it was previously taking.   By extending the analysis to include other data it is possible to have a more accurate understanding of the trends. Preventing out-of-stock situations

Big Data in Retail

Traditionally during the busiest times of year when a particular product was in high demand, the retailer who reacted to this demand fastest had the advantage.

Today the biggest retailers are using big data to gain competitive advantage. This is done by predicting trends and preparing for future demand starting with historical data analysis. fraXses deploys bespoke algorithms to support pattern recognition and machine learning capabilities as part of this historical analysis.

By combining enterprise data with other relevant data, such as web browsing patterns, social media sentiment, news, market analysis and research, it is possible to create predictive models of trends. When this is combined with location-based data, such as customer transactions, shopping patterns, local news, demographic data, local weather data and other location based research data, it is possible to prepare both the channels and individual stores with the relevant stock.

fraXses analytics provides insights used for price optimisation models and synchronising pricing based on demand, competitor activity and inventory. The personalisation of the retail offer is enhanced by combining customer data from loyalty programs, browsing patterns, social media, purchase history forums and demographic data. This customer segmentation allows for a much more personalised shopping experience with contact automatically made through their preferred channels and based on their current location with personalised real time offers.

fraXses also uses idea of the “what other shoppers like you bought” to make recommendations or create exclusive offers that will increase the average spend.

Retailers using fraXses provide a smarter shopping experience through:

  • Anticipating demand
  • Availability at right location
  • Dynamic pricing
  • Relevant promotions
  • Timely offers

With the objective of influencing purchasing decisions and beat the competition