About
Subscribe

Using historical patterns for future predictions

Lebo Mashiloane
By Lebo Mashiloane
Johannesburg, 04 Oct 2013

To stay ahead of their competitors, businesses need not only understand current consumer behaviour, but look at historical patterns to be able to predict future consumer behaviour.

This is the view of Goran Dragosavac, solutions manager of information management at SAS.

"Sometimes within large profitable segments, there may be pockets of customers who only look like profitable customers based on their market and behavioural characteristics when, in fact, their profitability is very low," explains Dragosavac.

Data mining technologies, he says, can be used to discover hidden patterns and apply them to predict future trends. In the retail environment, these technologies can help enterprises better understand the purchasing behaviours of customers - high and low margin customers, as well as help determine which customers are most likely to respond to a marketing campaign and identify which customers are likely to leave.

"The SAS Enterprise Miner is one of the ways to achieve this. It streamlines the mining process to create highly accurate analytics models based on large volumes of data from across the enterprise," Dragosavac adds.

According to Dragosavac, these models explore optimisation - which helps businesses determine what the best that can happen is; and predictive analysis - understanding what happens next. There's also statistical analysis, this examines why particular trends in the business keep occurring and forecasting, where future trends in product design, price and costs are taken into account.

To simplify this, Dragosavac uses an example of a business that sells umbrellas.

He elaborates that using analytics will help the business determine weather patterns - how many umbrellas are sold in a rainy or hot season; determine the quantity - how many umbrellas and models (wooden or plastic) to have in stock at any given time; and market segment - what kind of people buy these umbrellas and their geographical location.

"These factors will then inform the enterprise in structuring their marketing, determining which products and locations should be given priority due to their high profit margins. Also, by comparing historical responses with current ones, the enterprise can draft future projections, which may lead to identifying new opportunities for the business."

Dragosavac also touches on SAS Visual Analytics, which uses info graphs to help retailers of any size get insights through visual data exploration and flexible information sharing.

"This tool can identify elusive problems in the retailer's - vendor quality, error repetition, etc - by presenting visual graphs of supply chain data, sales transactions, call centre complaints etc.

He adds that it can enhance the consumer experience. Retailers can share analytical insights with store managers, employees and associates who can then use this information to offer personalised experiences to customers through preferred channels like mobile, social media or one-on-one e-mail correspondence.

"Innovation is a key component in retail. Mobile devices - which enable price comparisons, product research and online shopping - have put power literally in the hands of consumers," he notes.

"In response, companies should strive to provide flawless customer experiences and brand consistency across all available platforms, whether customers shop in a physical store, on a computer or on a smartphone."


Dragosavac concludes that retailers must understand customers and know their preferences well enough to anticipate their behaviours - and predict when those behaviours and preferences will change.

Share