We’ve identified 5 common challenges that retailers face that could be overcome with effective AI software and machine learning.
In recent weeks, the topic of stock returns has hit the headlines.
Clothing retailer H&M have followed brands such as Zara, Boohoo, Uniqlo, Dorothy Perkins and Next in charging customers for online returns. This has caused significant upset amongst customers, with many furiously declaring that they wouldn’t shop with retailers that charge for returns.
The topic is certainly contentious.
On the one hand, many shoppers are being forced to purchase online, either because of a lack of stock availability in-store, or because their local high street stores have closed down.
But retailers argue that dealing with returns is expensive. Not only is the time costly for warehouse staff dealing with stock returns, but they feel that the environmental impact is also something to be considered. Many brands, like H&M are charging a nominal fee to cover the costs of dealing with these returns.
But could AI be used by retailers to reduce stock wastage and prevent unnecessary returns?
We’ve identified five common challenges that could be overcome through the use of AI technology.
Challenge #1: Customer expectations
The news that more retailers are charging for returns has been met with criticism from consumers, who often state that an inconsistency in sizing or quality control are a key reason for product returns – especially for clothing brands. Many shoppers are selecting their preferred sizes, and finding that items are not as expected, even within the same store.
In fact, in an article published by the BBC, “H&M has admitted it needs to do more to improve consistency in clothes sizing.”
One of the ways that retailers can better set customer expectations is by including better communication on product sizes. Rather than relying on standardised size guides, individual items could have measurements specified which can educate the customer on which size to purchase.
AI is already being used to update product descriptions, so this could be an extension of the work that is already being undertaken.
We want to encourage retailers to invest in AI tools that are specifically designed to help a customer identify what products are suitable for them. Through a series of questions relating to their height, weight, body shape, age and gender, your AI could pull out key product recommendations that are likely to fit and suit the customer.
Not only would this aid personalisation strategies and increase the likelihood of a purchase, but it could reduce stock returns significantly by responding to the specific need to stabilise sizing.
Challenge #2: Fraudulent returns
Collectively, retailers are losing up to $1 billion per year as a result of fraudulent returns. Key issues include customers setting up new accounts to make use of promotional discount codes, fraudulently claiming that items have not been received, or returning significant number of items.
According to recent research from Riskified, “the start of the end-of-year holiday period is a key driver for specific types of abuse, such as returns abuse (38%), promo code and loyalty program abuse (35%), and item not received (INR) abuse (41%).”
Worryingly, the research has reported that:
Most merchants (62%) have no automated systems in place to address policy abuse. Integrating automated returns and refund systems could significantly reduce return volume, speed up returns processes, reduce errors, and thwart instances of abuse.
Preventing fraudulent returns is something that can be easily managed through the implementation of machine learning. Not only can it speed up the returns processes, allowing staff to have more time to investigate fraudulent incidents, but AI tools can monitor and track suspicious behaviour, particularly amongst repeat offenders.
With digital tracking technologies, retailers could easily identify where items are being purchased yet returned elsewhere – particularly through non-retail locations, such as parcel lockers.
Where required, your AI could identify the fraud rate on specific IP addresses, easily identify when an account was created or whether multiple accounts are set up for the same address or flag up concerns if a customer is making orders that do not match their purchasing history.
Challenge #3: Incorrectly labelled products
A key driver for stock returns is due to incorrectly labelled products. Manually managing SKUs is a tedious task with plenty of room for error.
But thanks to AI, a handheld device could physically check thousands of items on a product shelf in just a few minutes. Any mislabelled products could easily be identified and corrected – further reducing the likelihood of any stock wastage.
Another useful tool is to be able to use AI to check for food or drinks items that may be nearing the end of its shelf life. Your inventory management should allow you to identify which specific products are on the shelf.
If an item has a short shelf life, you can direct store staff to move those items to the front of the shelf and increase the likelihood of purchase. This could prevent the need to reduce the price of an item.
Challenge #4: Longer return periods
Historically, customers had only a week or two to ask for a return, which simplified matters for retailers, who knew that any sales/returns would take place within the same financial period.
But retailers are increasingly turning to longer return periods to entice customers, and to give them the time they need to make informed purchasing decisions. This means that they are often dealing with returns and financial losses over two or even three quarters, rather than one, making it much harder for effective financial management.
This is also creating a logistical challenge for retailers, as customers may be returning goods out of season, limiting any further potential for resale options. But AI could be used to support omnichannel solutions – perhaps by identifying any stores that are low on stock and need additional replenishment, or by flagging up specific locations of high demand for specific items.
This allows the retailers to ship their return stock to new locations, potentially boosting the resale value of that item.
Challenge #5: Staff management
One of the key issues with stock returns is the time that it takes for individual staff members to deal with individual returns. At busy shopping periods, this becomes even harder to manage (in-store and online) when retail staff are trying to prioritise sales orders, shipments and customer service.
Adding to the complication is the likelihood that busy shopping periods may also be a higher time for staff absences (e.g. winter colds, Christmas annual leave) – making it harder for individual stores to cope with a deluge of returns.
We recommend using AI to identify store locations with higher levels of stock returns, and cross referencing that with available staffing levels using your HR data. If there is not enough staff to manage the returns, AI tools can identify if there are any other locations that have more efficient staffing levels and reroute any returns to those stores.
Not only will this speed up the returns processes, but it could help you to get returned items back on the shelves quickly, improving the likelihood of a resale.
These are just a few challenges that can be overcome through the successful integration of AI tools and plugins. As technology changes, it's best to look holistically at the areas of your business that can be enhanced through AI software and machine learning.
By identifying where improvements are needed, we can recommend the right tools for the job, which will not only reduce stock wastage and prevent unnecessary returns but could also improve your bottom line.