Can you predict e-commerce returns? An analysis of 651.658 online orders and 412.584 returns

By Paazl
January 10, 2018

If December is the peak season for online retailer, then January is the peak season for returns. Unfortunately, record sales over the holiday season also means that e-commerce companies face a record number of returns at the start of each year. There is nothing you can do about it. Or can you?

According to a recent study by 3 researchers connected to the Germany University of Bamberg, there are certain signals that predict which products and what customers are more likely to return to sender. The team studied a sample of 651.658 orders placed at a German online store and the 412.584 (63%!) return shipments that followed. In terms of products 2.049.853 items were shipped out, of which 1.069.008 (52,1%) were returned. In total, this sample represents over €30 million in value – each order averaging €108.95. Lets take a look and see if we can predict e-commerce returns:

Variables that drive e-commerce returns

In the study, the researchers created two models using the dataset to discover which factors drive return rates, and to what extent.

1. The number of items in the basket

According to the study, ordering more items does not result in more items being shipped back. In other words, heavy shoppers do not automatically show higher returns rates.

2. Ordering multiple sizes of an item: the biggest red flag

As could be expected, ordering multiple sizes of the same style and color is the strongest signal that a customer is about to ship something back. This explains why certain online retailers, such as German shoe retailer Mirapodo, double-check with customers when they want to order multiple sizes. For more on preventing fit-related returns, be sure to read this article in which we take a look at 5 innovations such as virtual showrooms and fit-comparison tools.

A notification in the Mirapodo check-out reads: “Do you really need multiple sizes? Please note that each return costs money and pollutes the environment”.

3. Ordering multiple colors of an item: a good sign!

According to the study, customers that order multiple colors (of the same style and in the same size) typically return less. This could indicate that customers do not order multiple colors because they are unsure about which color to pick. More likely, they order multiple colors because they like a particular style, and want to own it in a wide range of different shades.

Are customers ordering multiple colors? Do not worry – more colors typically mean fewer returns.

4. Average price of an item: higher price = more returns

The likelihood of a return increases as the average value of each product increases. This could suggest that customers more thoroughly evaluate and inspect expensive items, to make sure they get value for the money spent. This also supports previous findings that show that customers returning more are typically the most profitable clients for an online store.

5. The use and value of discount vouchers. Give a little, get a lot

Customers using a discount code are more likely to return an item, perhaps because such vouchers trigger impulse purchases. The effect is small but significant – the use of voucher increases returns by 1,2%. This effect is reduced as the relative value of the discount increases, meaning a higher overall discount percentage reduces the chance of a return. In other words, as the feeling of having made ‘a good deal’ grows, customers are less likely to ship items back. This effect was already visible at a discount percentage of just 4%. So, if you are looking to increase sales and reduce returns, then consider the power of (low) discount vouchers.

6. Including freebies with orders: customers will ‘return the favor’, not the item

Similar to giving away discounts, giving away gifts also affects returns: the returns rate for customers who were given a freebie dropped an impressive 8,1 percent points! So seriously think about including small gifts or product samples if you are aiming to reduce returns. This could be explained by the psychological concept of reciprocity: people are likely to perform a positive action in response to a positive action by somebody else.

A good example of reciprocity being used in e-commerce comes from Canadian company Gameklip, who include candy with each order, leading to very positive reviews.

7. Using a post-delivery payment method

In recent years we have seen the rise of post delivery payment methods, allowing customers to pay after their items have arrived. These payment methods are a proven tool for increasing conversion. But according to the study, this method might have a shadow side. Customers using post delivery payment return up to 10 percentage points(!) more, compared to those using regular payment methods. This suggests that customers who ‘buy before actually spending’ are potentially very expensive for online retailers. To combat this, online retailers might consider disabling post-delivery payment methods for those customers who have a history of frequently returning items.

Paying after receiving your goods is great in terms of customer convenience, but might increase the risk of returns.

8. Previous customer behavior

Speaking of previous behavior: according to the study, previous returns behavior by customers is a good predictor of their future returns rates. This supports the existence of so-called serial returners who show excessive returns behavior. To keep track of these costly customers, retailers can subscribe to databases that keep track of notorious returners. While it is legally not allowed to ‘ban’ returns, there have even been rare incidents of retailers banning customers for returning too much.

So, what can you do?

Even though this study only covers the returns at a single online store, it offers some valuable insight. Especially, because of the large sample size. These are the actionable steps you can take:

• Double-check with customers who order multiple sizes of the same item. Is it a mistake? Do they really need it?

• Use discount vouchers to reduce returns. As little as 4% discount can lead to a decrease in returns.

• A small gift goes a long way: the study found an 8.1% drop in returns for orders that included a freebie.

• Carefully look at payment via invoice. While this might boost sales, it is also connected to a rise in returns. Consider disabling this payment method for those who often return items.

• Keep track of your serial returners – outright banning might not be a solution, but there are other ways to take action such as a personal call or a stricter returns policy.

Looking for more e-commerce delivery inspiration?

Be sure to discover our latest whitepaper in which we analyze the delivery & returns policy at 26 online fashion stores:

More inspiration.

See Paazl in action.


Get the Paazl picture.