Home Retail The rise of predictive analytics in retail

The rise of predictive analytics in retail


1Leveraging Data

The best way to predict the future is to create it. And a few leading retailers are doing just that. They are leveraging ‘data’ and ‘analytics’ to inform better consumer, product, pricing, marketing, and supply chain decisions among others. But what’s really fueling this?

Customers search for products, check product prices, read reviews, refer to ratings, and even share their preferences via social media

The rise of the digital consumer, hyper competition, and technology are fueling this unprecedented change. Over 70 per cent of shoppers begin or use the internet to inform their purchases. They search for products, check product prices, read reviews, refer to ratings, and even share their preferences via social media. As a result, they are leaving behind a huge trail of digital intent signals. There has been a significant growth in the number of e-commerce players and models. While there will be consolidation and sanity, it will very likely change the expectations of the shoppers for good. And finally, the rise of big data, cloud computing, and mobile connectivity.

So why predictive analytics, and where is it being applied? I shall illustrate this with three examples across retail viz. customer loyalty, pricing, and assortment decisions.

2Knowing Me Better Can Be Rewarding

Customer loyalty has been an area where a lot of predictive analytics have been at play especially around trying to predict which set of customers will churn. But what’s changing in this new paradigm is the enrichment of the churn models with these newer sources of data.

Google APIs now allow one to understand distance of the customer’s address from the most frequented store and from that of the competition. Social signals enable one to understand preferences. Web log data helps understand web behaviour. A lot of these data sources are still not being used in churn models by most retailers. But as store retailers begin to establish or strengthen their multi-channel plans, there will be a great opportunity in improving their loyalty analysis based on these newer signals. There is evidence that predictive analytics based on these signals results in retailers being more relevant to their customers resulting in better business performance.

3The Right Price Is Right Now

Dynamic pricing is a reality in most categories. Shoppers have embraced this new paradigm. But, most retailers have not. There are a leading few who have built sophisticated predictive algorithms to determine what price the customer is willing to pay. They analyse demand signals, supply, competitive information, and a whole host of other parameters to arrive at the right price – right now.

The more advanced retailers are also modeling competitive price elasticity into their predictive techniques to make sure they are not just playing the price match game.

4Will This Sell?

The rise of marketplaces has created enormous choice for shoppers. Complement that with super flexible supply chains and you have the perfect recipe to pamper shoppers with choice or make them fickle. Merchandizing teams at retailers can no longer rely on traditional approaches to assortment selection where they rely on past transactional data, anecdotal information from suppliers, and third-party market share data.

The enormous growth in consumer demand signals (like search, reviews, rating, social likes and pins) are now being leveraged by leading retailers. They are incorporating them into predictive algorithms to guide decisions on what products to keep, carry, and drop.

We are at an early stage here and are just scratching the surface. Predictive analytics will become mainstream in many decisions in the form of on-going nudges. That’s not hard to predict.

ABOUT THE AUTHOR: Mihir Kittur is the Co-founder and Chief Innovation Officer at Ugam. He holds a B.S. in Electronics and Telecommunications Engineering and a Master’s degree in Business Management.