How A New Approach To Customer Segmentation Can Reduce Returns And Boost Profitability Up To 30%

0aaaSimon Hay Outra

Customer returns are reaching pandemic proportions. Increasingly, shoppers are being conditioned to return unwanted items, ‘nearly unworn’ or over-ordered clothes in a sector with already tight margins. Through a new approach to predictive customer segmentation, now retailers can very quickly and cost-effectively identify those customers most likely to erode profitability. By using this insight and adapting communication strategies accordingly, it is possible to significantly reduce the cost of returns on the P&L — and maybe help a bit on CO2 emissions too!

The Emergence Of A New Group Of Consumers

In the UK, the cost of customer returns ever year is the same as the BREXIT divorce bill, a not insignificant £60bn2, whilst in the U.S. this amounts to $351 billion1. And according to a study by GlobalData the return rate is expected to increase by a quarter over the next five years. Returns are now the new normal and online shoppers are increasingly being encouraged to buy on the understanding that they can return. Zappos started the trend of an attractive returns policy to drive sales by offering a 365-day, free two-way shipping and returns service that invited customers to order its shoes, try them on at home, and send them back if not completely happy.


A new school of thought quickly emerged extolling the practice of simple returns to both achieve competitive advantage as well as provide a healthy source of customer loyalty and profits. This has led to the emergence of a new group of customers; those that buy one clothing item in a range of sizes and colours, treating their home as the fitting room whilst knowing in advance that they will return at least one. And even if they miss the return date, there is a healthy online market for new clothes complete with tags. This behaviour is further being encouraged with services such as Amazon Prime Wardrobe and Nordstrom Trunk Club. However, the reality is that most brands simply cannot afford to entertain this shopper behavior.

A New Approach To Customer Segmentation – One Size Does Not Fit All

Traditionally retailers have benchmarked returns against industry figures; however, this can be significantly skewed by brands such as Zappos. In this case, Zappos identified that customers purchasing their most expensive trainer also had the highest return rate of 50%. Clearly curtailing the returns offer now will have major short-term impacts.

It is necessary to take an individualized approach to provide insight into how a brand’s own customers are behaving. This can be achieved through ‘deep learning customer segmentation,’ a form of AI that is proven to be far more accurate in predicting customer behaviour. This can be used to determine how to manage different customers according to their propensity to return and other criteria we can use to optimize customer value — short- and long-term.

A customer returns segmentation is a powerful tool to help businesses identify their most and least profitable customers, taking into account the rate of returns. This can then be layered against additional insight such as demographics, RFM segments, LTV models, category behavior and time trends to provide a fuller understanding of the customer and their behavior in real time.

Identifying a customer’s returning habits along with in-depth transactional information allows businesses to understand a customer’s true value and allow retailers to be more relevant across the whole customer experience. Once these segments are identified, they are broken down into a more granular level and overlaid with social media usage to recognize which customers actively encourage this behavior amongst their peers. This insight is used to activate automated responses to ensure that opportunities to engage with the most profitable group of customers are not missed and to ensure the least profitable customers are not sent offers that ultimately lead to an unprofitable action.

Segmentation can also run deeper than communications; it can be used to inform logistics policy. For example, establishing a personalized returns process where the best customers are offered free shipping and returns, whereas less profitable customers have to pay for it.

Returns have become fully accepted as a cost of doing business online, but unchecked and ever-growing levels of returns do not need to be the case. A study by The Reverse Logistics Association reveals that reducing returns can improve profitability by 30%. With the advent of more insightful and predictive segmentation it is possible to predict customer behavior and create a tailored approach that significantly impacts business performance.


Simon Hay is the CEO of Outra, which provides automated and actionable customer insight to retail brands powered by artificial intelligence.

1 GlobalWebIndex


3 GlobalwebIndex

4 Invesp

Feature Your Byline

Submit an Executive ViewPoints.


Access The Media Kit


Access Our Editorial Calendar

If you are downloading this on behalf of a client, please provide the company name and website information below: