Around 7 in 10 people abandon their cart during checkout, according to a meta-analysis by Baymard Institute.
The good news is that up to 15% of these lost customers can be recovered with data-led interventions, making abandoned cart recovery a fundamental use case in the world of e-commerce.
How does abandoned cart recovery work?
1. A user adds items to their cart
At this stage, you need at least a user identifier and the product identifier (SKU) for their basket.
The user data is often modeled from a larger data set. When this data is grouped by “user id”, you can see all the information you have pertaining to that user, such as their actions, purchases, demographic information and so on.
2. The user does not complete the purchase within a certain time period
You’ve defined what constitutes an abandoned cart and these criteria have been met – maybe 12 hours have elapsed and still no checkout.
The data you have available now dictates your best course of action. If you are using behavioral data, you can look at the actions the user took leading up to and possibly after the abandoned cart (while they are using your sites or apps). The more information you have about a given user, the more experiments you can run to see which behaviors are most indicative of a high-value customer to re-engage.
If you are working simply with transactional data, your responses will be very limited and based on the only information you have available – e.g., purchased (y/n), email address, etc. Clearly, not knowing the actions a user has taken on your site is a significant disadvantage when it comes to reengaging them.
3. You contact or nudge this customer through one of your channels in order to encourage them to complete the checkout process
Working out which channel is optimal for which customer is important. A text, an email or an in-app nudge are all effective for the right customer.
Another consideration is timing, experiments can be run by grouping your customers into segments to establish the optimal times to re-engage.
It may not be worth attempting to recover all abandoned carts; if a user is identified as low-value, the cost of this process could outweigh the benefits (e.g., a $2 cart abandoner). This requires more advanced segmentation.
4. You measure the success of your recovery campaigns and tweak them accordingly
Creating cohorts of users allows you to create control groups and run experiments to measure the success of your assumptions.
It may be that one tactic works for users who were ordering high-ticket items, while it’s another for more everyday products. You can keep tweaking the logic and wording of your campaigns if you have the right data to measure the success of your interventions.
“$4 trillion worth of merchandise will be abandoned in online shopping carts this year, with about 63% of that potentially recoverable by online retailers"
Marketplace Manager
What is a good abandoned cart recovery rate?
This question be broken down in stages with the following rough averages:
a) The number of abandon cart emails that are actually opened:
~50% on average.
b) The number of these email openers who go on to click through:
~20% on average.
c) The number of those that click through who go on to convert:
~10% or more is considered good
The challenges with abandoned cart recovery
1. Abandoned cart emails and GDPR
Article 4.11 of the GDPR states that consent must be “free, specific, informed and unambiguous”, and that any infringement of this rule may result in fines of up to €20 million or 4% of annual revenue, whichever is larger.
Understanding how GDPR affects abandoned cart emails will help you stay compliant and prevent paying hefty fines. This can be a real challenge, depending on the processes you have around data collection.
Learn more about GDPR and abandoned cart emails
2. Finding the balance
If you can accurately predict Customer Lifetime Value (CLTV), based on certain behavioral profiles, you can segment customers into different groups based on potential value. This advanced segmentation means you can focus on iterating on your communications with high-value prospects.
The caveat is that getting an accurate predicted CLTV requires rich and high-quality data, modeled effectively so as to be useful for your data teams. Many find this to be a difficult use case to achieve due to the number of data points which feed into your final figures. This comes down to the richness, organization, and quality of your data.
The right time:
There is a fine line between re-engaging and annoying potential customers. The wrong word at the wrong time means lost sales opportunities, so fully understanding the reasons a customer abandoned their basket as well as conducting effective experimentation is important.
The right words:
Choosing the best copy to resonate with your target audience is fundamental to cart recovery, but without high-quality data, it is often guesswork. With basic data, you can say that a certain email got a better CTR or open rate, but you can’t define which copy will work with which users in which circumstances. This requires digging deeper by using advanced analytics.
The right channel:
Finding the right channel for a given group of users requires organized experimentation followed by automation based on behavioral profiles. The advanced analytics or AI needed to achieve this presents a challenge in terms of technical sophistication.
3. Deciding what actually counts as abandoning a cart
You need to think about the precise definition of cart abandonment and decide whether this works for your business. Is it:
– Not returning to a shopping cart for 15 minutes?
– Leaving your site or app after an hour?
– Still not having made the purchase by close of play?
– And so on…
You then need to test this assumption across your segments, these steps present strategic and technical challenges.
Abandon cart recovery with Snowplow
Snowplow is a Behavioral Data Platform, built specifically for capturing the smallest details of user behavior and sending this data to your storage destination in a highly-usable format.
This allows you to take your abandoned cart projects to the next level by freeing your data team from endless wrangling and preparation, allowing them to conduct sophisticated experiments to maximize re-engagement.
Snowplow BDP can be integrated throughout your e-commerce store and checkout flow to capture incredibly rich data on your shopper behavior. This data is delivered in real-time, meaning that abandoned carts can be quickly detected and reacted to, boosting recovery rates.
Here are some of the features of Snowplow which help you to reengage more customers:
Robust user identifiers
Snowplow provides out-of-the-box user IDs, including session IDs, cookie IDs, and IP addresses.
First-party server-set cookies
These cookies can track customers for up to 400 days (at the time of writing), including Safari users normally blocked by ITP restrictions. Snowplow is actually deployed in your own cloud environment, ensuring this is done in a highly compliant way.
Out-of-the-box e-commerce features
Snowplow provides a set of e-commerce tracking methods and associated JSON schemas that you can use out of the box (or you can create your own custom data structures for your own specific business needs).
130+ metrics tracked automatically
Other metrics which can help you optimize abandoned cart recovery, such as the page the user was on before and after the event happened, the device, location, and other specific user information. Take a look at our modeled data.
Event properties are kept within a single table
This means generating user segments to answer advanced questions is straightforward. These questions include the optimal time to send communications, what counts as an abandoned basket, and which channels to choose.