Product Data and Returns: Causation or Correlation?

The difference between causation and correlation is often misunderstood when applied to product returns. Did the bad product data cause the return, or did bad product data correlate to increased product returns? In a rudimentary sense, all causations involve correlation, but this post is about looking deeper into the reasons people return products and deciphering the cause from the effect.

First, we need to look at why people return products, both from an ecommerce and brick-and-mortar business perspective. Then we will examine the root cause to determine if product data causes those return justifications, correlates to them, or has no impact at all.

The Reasons for Product Returns

First off, we need to agree on a few facts. The first is that 17.6% of ecommerce sales as a percentage of dollars sold end up being returned, as opposed to 10% of brick-and-mortar sales (NRF, 2023). This makes returns a $743 billion USD per year problem, and the problem is growing at over 20% per year. Return fraud alone now accounts to $100 billion USD (ChargeBacks911.com, 2024), and some products, like apparel, have a 50% return rate.

These numbers are staggering, especially when you consider the costs of returns to a business. They range from shipping and re-packaging costs, warehouse storage costs, resources to manage returns, and product disposal costs that go all the way to our local and global landfills. Tackling returns can lead to large gross profit gains that are unlikely to be solved with market share gains or cost-cutting alone.

So why do people return products at such alarmingly high rates? The top 5 reasons for returns according to multiple online sources include:

• products damaged in shipping
• the wrong item being shipped,
• the item looking different than what was shown on the website
• customers changing their minds after receiving the item
• “wardrobing”.

We will look at each of these individually to see if there is a correlation or causation from the return justification to the product data.

Damaged in Shipping

Items damaged during shipping make up a high percentage of returns. This may seem to be totally independent of product data, but the truth is it is not independent. Not marking products as fragile, sending through the incorrect shipping mechanisms, and failing to perform the analytics on what products and shipping methods create the most damaged items both cause and allow damage during shipping to continue. Some damage will always occur, but product data can assist in reducing the rates of damage. Therefore, a lack of product data quality can be a cause of some product returns.

This means managing your product data has to include post-sale data as well. Most businesses see the post-sale data as a product development issue, or they do not look at it at all. This neglect in managing this data can be a significant cost, as over 50% of B2C customers stated they returned an item damaged in shipping.

Wrong Item Shipped(?)

We have all at one time opened a box from an ecommerce site only to look at what is in that box with dismay. Finding the wrong item in a box is not only disappointing: it is way too commonplace. But does product data play a role in that disappointment? The answer is: Maybe.

First off, we need to understand that sometimes what we get in the box was actually what we ordered. We just did not like what we ordered, or it didn’t look like what we expected. Sometimes it is absolutely the wrong product, but sometimes it isn’t. Although 45% of consumers say they have received the wrong item, another 30% stated the item they received didn’t look like or act like what they saw online. This makes this return a reason a huge problem for retailers.

Item looks or Performs Different than expected

Product data cannot help make sure the right physical item was placed in the right picking location in a warehouse, or that the packer put the right label on the box. These are not causation or correlation issues, and are more likely inventory control problems. Even ensuring the right color of apparel is sent to the right customer involves process issues that are mostly abstracted away from product data.

But the wrong picture being shown online, the specifications not being accurate, or not posting enough data for the customer to make an informed decision are all product data issues that cause returns. Improving data accuracy and completeness is vital to avoiding customer returns. Instead of looking at how to facilitate easier returns, companies must understand how to keep these products in the hands of the people they sold them to. This is done by selling them the right item the first time. Selling that customer the right item the second time bleeds margins, while leaving potentially unsellable products in their wake.

Customer Changed Their Mind

There are three reasons customers change their minds after making a purchase. The first is they regret making the purchase, either due to the quality of the product or their own expectations. The second is that they found a different solution while waiting for the product to arrive, changing their need sets. The third is that their financial situation changes, either due to unexpected expenses or they never could really afford it in the first place.

Very little of this can be solved through product data. Better images, specifications, and other data may be able to avoid a purchase that does not meet expectations, but financial concerns and customers no longer needing the product cannot be solved by better product data.

What can be solved is some of the fraud that is occurring on a daily basis in the ecommerce space. This is a variant of this return justification: The customer never intended on keeping the product. Sophisticated fraudsters will only be trackable based on tracking down those people when they are identified after multiple returns. They have made a living by understanding the rules and processed involved in returns, and will even invest money in their schemes. There is very little product data can do to stop these types of retail crimes.

However, the average fraudster is not that sophisticated.  They open the box, replace the item in the box, try to seal the box, and then ship it back. Their hopes are that they already have the purchase price returned to them before somebody finds out what they did. But can product data assist in detecting these issues? Maybe, and only in relation to good returns processing.

If the resource processing the return has access to good product data, like weights, dimensions, colors, and specifications, they can identify a fraudulent return earlier in the process. The item does not end up in someone else’s hands before the fraud is realized, saving a further negative customer event. Good packaging information on electronics assists with higher impact, but detecting fraud can come from good product data.

This is one of the rare instances of a correlation between product data and returns. The less product data you provide, both at the time of sale and at the time of return, the more likely the culprit is to get away with their fraud. Product data didn’t cause the fraud to occur, but it can assist with your detection of the fraud.

“Wardrobing”

Wardrobing in this case is a variety of intention return types. True wardrobing has been around for decades, where someone buys an article of clothing and keeps the tags on it so they can return it after a special event. In electronics, showrooming became a big issue from a sales perspective. Showrooming II became a bigger returns issue, as people would purchase from a local store only to return it after receiving the same item at a cheaper price from an online store.

Product data can do little to avoid this type of return. The intent in the purchase was never to keep the purchase, and beyond good analytics on customer return data, product data cannot avoid this type of return. There is no causation or correlation between the categories of products likely to incur this type of behavior. It is a form of fraud that requires a different process or return policy response.

How to Lower the Cost of Returns Through Product Data

The cost of attracting one more sale is expensive. If that sale results in a return, that cost is even higher. Gaining 1% more market share becomes even more expensive when returns outstrip the rate of sales growth.

The easiest way to reduce those costs is to keep the product the customer bought the first time in their hands. Whether that’s through better identification of shipping methods by product to avoid damage, more complete imagery and specifications to avoid incorrect purchases, or guided selling to ensure the customer makes the correct choice on the product that fits their needs, product data plays a role in this aspect of returns. Investing small amounts in how data is collected, how complete that data is, and making that data available to both the consumer and in the returns process can lead to a sharp decrease in returns and big gains in maintaining margins.

Bibliography

ChargeBacks911.com. (2024). Retrieved from The State of Product Returns in 2024: https://chargebacks911.com/customers-returns/

NRF. (2023). NRF and Appriss Retail Report: $743 Billion in Merchandise Returned in 2023. Retrieved from https://nrf.com/media-center/press-releases/nrf-and-appriss-retail-report-743-billion-merchandise-returned-2023#:~:text=Online%20sales%20do%20see%20a,store)%2C%20or%20%24371%20billion.