Optimising retail operations with the latest technologies

Microsoft’s ShiSh Shridhar explores how advanced analytics, machine learning and the cloud are driving optimisation and accurate demand-driven inventory and assortment planning

Rebecca Gibson
Rebecca Gibson
By Rebecca Gibson on 12 October 2016
Optimising retail operations with the latest technologies

This article first appeared in the Autumn 2016 issue of The Record.

Understanding what customers want, how much they want to pay and when they want to buy products are key considerations for retailers making inventory decisions, says ShiSh Shridhar, Microsoft’s worldwide director for business intelligence solutions.

“If retailers or consumer packaged goods (CPGs) companies are regularly experiencing out-of-stock situations on popular products and overstocking on others that aren’t selling well, then they’re clearly not in tune with what consumers want and this will eventually cause them to lose customers, sales and profit,” he comments. “Factors like the weather and the economy, as well as the age, gender, education level and socio-economic status of customers all have an impact on which products are most popular in stores at particular times of the year. It’s critical that retailers understand how these different elements impact their sales so they have the right number of the right products in their stores at the right time.”

According to Shridhar, advanced analytics and machine learning allow retailers and CPGs to get real insights that enable them to quickly and easily predict consumer buying habits and deliver an in-store and online product assortment to match.

“Traditionally, retailers had to rely on historic sales and customer data to predict what products would be most popular with their customers, but as customer preferences and market trends change quickly, these forecasts were not always accurate,” he remarks. “Today, they can use advanced analytics to capture unstructured data from various sources, including social media and customer review sites, and mine it intelligently to find patterns and make predictions that will ultimately help them to resolve issues or optimise operations.”

Microsoft’s cloud-based Microsoft Advanced Planning Solution enables buyers in retail and CPGs to combine publicly available unstructured data with their own structured operational and transaction data. Advanced analytics and data visualisation are powered by Microsoft’s Cortana Intelligence Suite, Power BI and Azure Machine Learning.

According to Shridhar, the solution helps users to optimise resources for demand forecasting, inventory and assortment planning, churn predictions, price optimisation and much more.

“Cortana Intelligence Suite, for example, allows retailers to pull streaming data from their POS devices so they can monitor product sales during promotional campaigns,” he says. “This information can be combined with customer footfall data captured by in-store sensors and internet of things devices so retailers can make real-time updates to the offer, target a new demographic, or move the products to an area of the store where customer footfall is significantly higher. Sensors are also optimising the supply chain by allowing retailers to quickly re-route products according to changing customer demand in different locations.”

To date, Microsoft and partners like Cognizant, K3 Retail and 11Ants Analytics empower retailers and CPGs to optimise operations with machine learning and advanced analytics. For example, Microsoft and Neal Analytics helped Mexico-based beverage manufacturer and distributor Arca Continental – also the second-largest bottler of Coca-Cola in Latin America – to make its demand forecasting operations more accurate.

“Using Azure Machine Learning we identified the factors that led to people buying Coca-Cola across Mexico, finding that consumption spiked on hotter days and when the city was hosting an event,” says Shridhar. “We ran all of this data against Arca Continental’s historic sales data so it could accurately predict demand across stores in various cities at different times of the year and reduce overstock and out-of-stock situations, thereby improving profit margins. The company also used our platform to envisage the impact increasing investments in areas like advertising, or making pricing more competitive, would have on its total sales and profit margins.”

Similarly, US-based home décor and furniture retailer Pier 1 Imports deployed a predictive analytics solution based on Microsoft Azure Machine Learning and Power BI to better plan product assortments and inventory at an individual store level.

“The solution combines online and in-store transactional and behavioural data to determine what type, colour and style of products are most popular with customers in certain towns or cities,” says Shridhar. “For example, if Pier 1 can see that in winter, Seattle-based customers look for rugs while those in Miami buy more cushions, it can ensure that product assortments and stock levels reflect this demand.”

Certainly, notes Shridhar, machine learning and advanced analytics, are unlocking new opportunities for retailers to transform operations. “Now they can envisage what might happen and predict customer demand, they can understand how to sell more products and ensure the right level of inventory and assortment is available, creating a personalised and seamless customer experience.”


Number of views (3301)/Comments (-)

Tags: retail
Comments are only visible to subscribers.

Theme picker