Predicting future outcomes in an unpredictable industry
By Archel Aguilar, Managing Director, StayinFront Group Australia Pty Ltd
The consumer goods industry is anything but predictable, especially in the alcoholic beverages sector.
Things can change in an instant for any given reason. Beer, wine, spirits or the new trend – alcoholic seltzer products, could be discontinued, stock could go missing from the shelves or a million other factors could cause a less than optimal customer experience.
Even further, unpredictability comes from the lack of brand loyalty in Australia, with only one in four Australian consumers exclusive to a specific liquor category.
Getting the right brands and products in the right locations is vital, and so is getting the appropriate service model for a diverse range of outlets.
Driving the correct in-store actions in a busy category like the drinks industry - with a proliferation of competing SKUs - requires effective tools that can help companies achieve their selling goals - even if they do not have resources like Electronic Point of Sale (EPoS)/scan data.
Historically, off-premise tools used to segment call files, and required store level sales data to inform the algorithms.
However, driving retail execution with on-premise sites in Australia is complicated, with a diverse range of outlets and a rapid increase of brands competing for space and airtime. Trying to figure out which stores need the most attention from a field rep and which don’t can be a daunting task.
From big-box supermarkets to independents - which still represent 10% of category sales - store operations vary. EPoS data covers 65% of the liquor market, and the remaining 35% can use predictive analytics.
What are predictive analytics?
Predictive analytics models use a wide variety of data sources to estimate the likely sales value in any given outlet, despite the absence of sales data across the outlet universe. These models use intelligent algorithms to identify, from a wide variety of store characteristics and variables, those that have the closest relationship to sales, building a store level forecast that segments stores according to their estimated sales.
By supplementing the supplier’s view of a market’s store universe with the actual store universe, predictive models can also investigate stores the supplier may be unaware of. This minimises the risk that the bottom performing 10-20% of the stores currently called on by the sales force may represent a lower sales potential than the top 10-20% of the stores not currently visited at all.
Armed with a reliable forecast of sales per outlet, suppliers can construct a contact strategy for their sales forces that closely mirrors the approach taken in markets with EPoS data.
StayinFront’s 20:20 RDI Predictive Analytics supercharges field sales by using a wide variety of data sources to estimate the likely sales value in any given outlet, despite the absence of scan data across the outlet universe. With limited data, new PA models can answer the burning questions of “where should I deploy my resources?” and “which brands/SKUs are most appropriate for the outlets catchment area?” Combining artificial intelligence (AI) to turbocharge a supplier’s commerce with machine learning algorithms can revolutionise the focus for field sales.
For more information, read StayinFront’s White Paper on Predictive Analytics here.
Archel Aguilar is responsible for the Australia and New Zealand markets and operations throughout the Asia Pacific region. He joined the team in 2003 with extensive experience in the computer software industry. His skills and experiences working with Business Processes, Requirements Gathering and Analysis, Agile Methodologies, Project Management and Customer Relationship Management have helped him work his way through the ranks.
StayinFront is a Corporate Partner of The Drinks Association.