Algorithmic Retailing

Taming the Returns Monster with AI

 
January 24, 2020

If location, location, location is the success mantra for brick and mortar retail, then convenience is the mantra for online retail. As retailers pull no punches to make the transition to online shopping easier and lucrative for consumers, ecommerce is expected to become the largest retail channel in the world by 2021, outpacing sales through retail outlets, like supermarkets, independent grocers, and apparel and footwear retailers, among others, according to a February 2019 report by Research and Markets.1

Moreover, retailers are being lured online by the prospect of lower cost to serve and the promise of reaching a large customer base. However, the growth of online retail is not all roses. Conveniences such as hassle free returns have now turned into a fire spitting monster threatening to scorch retailers’ profit margins. GlobalData has forecasted online returns to increase to USD 7.2 billion (£5.6 billion) over the next 5 years.2

The financial impact of returned goods can be overwhelming at several levels:

  • Involves high reverse logistics costs
  • Processing and refurbishing returned items costs between USD 6 and USD 20 per returned item3
  • High indirect costs such as customer care
  • Cost of missed sales while it’s being processed for return
  • Unsold inventory costs, especially in the case of fashion and seasonal merchandise
  • Environmental impact as dead stock that is unable to be sold at full or discounted price finally ends up in landfills causing a huge carbon footprint

The mounting cost of returns is a monstrous problem for many retailers that simply cannot go unchecked much longer. Hence, the need for predicting returns and being more pre-emptive.

AI-powered anticipatory returns: The retailer’s genie

Modern retailing needs modern solutions. New age technologies combined with artificial intelligence (AI) and machine learning (ML) can help retailers come to grips with returns in the following ways

Predict returns before order placement

AI enables real-time assessment of the customer’s basket correlated with the customer’s individual profile based on preferences like fit, order value, and purchase history to predict the probability of the product being returned with a great degree of accuracy. AI can check for:

  • How many similar items are there in the cart
  • Are the items of the right fit
  • What has the customer's return rate been for similar items in the past

These insights can be extended to an AI powered decision-science engine that can make recommendations to prevent a return outcome. Recommendations include offering a targeted promotion, suggesting better suited substitutes, free delivery charges, or even dissuading serial returners from making the purchase. For example, it may increase shipping charges as a deterrent or offer a voucher as an incentive in return for making the purchase non-returnable.

Forecast returns after order placement

AI can build a contextual assessment of captured orders, customer profile, purchase history, credit scores, social media trends and product performance from historical and current data to forecast returns. This helps retailers to plan for reverse-logistics in advance, lowering costs for retailers. AI can also help identify and manage errant customers who abuse return rules by ‘wardrobing’.

Link return rate to other business areas

AI can turn returns into opportunities and bring efficiencies across the value chain:

  • Sourcing: Intelligently link suppliers to product return rates with explainable AI to ensure right products at the right price in the future
  • Marketing: Customize targeted marketing based on customer returns profile
  • Supply Chain and Logistics: Deploy AI/ML to pre-empt returns inbound for better depot and resource management
  • Merchandising: Leverage the power of AI to optimize the ranges to reduce returns resulting from issues such as fit and size. Right range at right price could mean higher customer satisfaction.


Eliminate reasons for returns

By assimilating, assessing, and drawing contextual insights from internal and external data sources, AI can help retailers understand customers better. This knowledge can be used to ensure that the customer buys with confidence and their key reasons for returns are eliminated:

  • Make personalized recommendations: Leading fashion and sports brands are using AI to provide customers with customized recommendations when shopping online. For instance, customers can interact with Levi’s virtual stylist using Facebook Messenger to find the perfect pair of jeans.
  • Determine fit: Size and fit are the primary deal-breakers. AI helps shoppers make informed decisions through multi-point sizing that delivers a bespoke fit. ML and AI enables brand-to-brand size translations, leading to better fit recommendations for online buyers. UK’s top fashion retailer ASOS uses AI to make size recommendations based on the customer’s purchase history.
  • Predict price points: Tools like TCS’ Optumera™ apply real-time computational intelligence to track and pre-empt competitor's prices, assortment and inventory across channels to enable retailers to create winning competitive strategies and reduce returns.
  • Make accurate delivery promises: In case of next day delivery, AI can have a holistic view of inventory positions in real time to help retailers keep their fulfilment promise. Not only will this curb returns but also it will do wonders for boosting customer confidence and loyalty.


Returns Management: No smooth sailing

Despite different options available for managing returns, there are a few inherent challenges as well: 

  • One option to reduce the amount of returns is to enforce strict policies around product returns but that could be detrimental for the business. As quoted by a leading chair brand, "If you make it harder at the front end then people will purchase less but may return just as much. Returns must be looked at as another cost of bringing sales online. You don’t have an option – the public know they have all the cards, so they will vote with whoever they think is being the fairest.”
  • Product purchase history and return records contain rich information but can be challenging to integrate in a principled way for the purpose of predicting future returns. Moreover, historical data points are not sufficient as the returns are dependent on several factors.
  • Retailers feel their physical stores can actually be assets and are encouraging online shoppers to return/exchange products in their brick-and-mortar locations. This is based on the assumption that consumers' return rate within stores is less than online. But, this does not work for a pure-play online retailer and also is a step back from the digital world retailers are embracing.
  • While some brands have the financial wherewithal to absorb the hits—or will mitigate the costs in a way that does not materially impact the customer experience—most cannot.


Returns are the new norm 

Retailers are slowing waking up to realize that instead of fighting returns head on, a more bottom-line driven approach will help business. Moreover, accepting returns as central to customer experience can help retailers deal with it differently and it may cease to be a problem. In fact, how you deal with returns—pre and post purchase—can bring brand differentiation, cut competition, and even make you more profitable. With the right set of tools, returns can be a source of loyalty and growth.

Sanjeev Jha is Head - Retail Strategic Initiatives (UK/Europe), TCS. As a certified AI & Management Consultant at TCS Retail, Sanjeev Jha is a key member of the Retail Strategic Initiatives group, a niche group that drives growth and transformation strategies for leading retailers within the European geography.