In our prior blog post we created a simple numerical example that demonstrates how improving CAR at the bottom of the web sales funnel can have 2x the impact versus increasing the conversion rate at the top of the funnel.

The purpose of that example was to get your attention! The purpose of this post is to guide you on how to further analyze your CAR metric, so you can properly prioritize your e-Commerce efforts.


The simplest way to evaluate your own CAR metric is to benchmark it against the metric used by industry peers.

In spite of frequent press about cart abandonment rate, there is a total shortage of reliable comparative data. A few companies, like Digital River) make benchmark data publicly available. However, in general, you are left to rely on your e-Commerce platform or Web Analytics provider to share their aggregated benchmarking data with you.

Therefore, applying anecdotal data points to your own site performance is not a recommended approach to your CAR evaluation. Furthermore, no company should be intimidated by significant cart abandonment rate because we still don’t fully understand all the factors that impact the abandonment and what should be an acceptable rate for different types of business.

What we do know is that the abandonment is a part of the normal buying cycle and it is not necessarily an indicator of the brand rejection or poor checkout process. Many visitors will return multiple times as they consider the purchase, while storing items in the shopping cart as ‘wish lists.’

CAR metric variations

As we already mentioned in the prior post, CAR is the opposite of the cart conversion rate (CCR, percentage of all orders that ended as fully completed buying transaction):


Note: we used CAR=CAR(O)= 1-CCR(O) notation to indicate that the metric is based on the number of orders and purchases.

CAR(O) metric does not directly correlate to the revenue trends nor does it paint a full picture about the abandonment details. This is why we also recommend  use of variations of this metric, that show the abandonment rate based on the value of all orders and purchases, or the ratio between the number of items in the orders and purchases.

CAR(R) – CAR for Revenue: This metric should be calculated on the basis of the total value of all items placed in the cart and total value of all completed purchases. This metric shows the % of the revenue that was abandoned.

PURCHASES – total value of all purchases
ORDERS – total dollar value of all Add-to-Cart items

CAR(I) – CAR for Items: This metric should be calculated on the basis of the total value of all items placed in the cart and the total value of all completed orders. This metric shows the % of items that were abandoned.


#PURCHASED_ITEMS – total number of items included in all purchases
#ORDERS_ITEMS – total number of items that were added to the cart

Smoke Tests

Focusing on specifics of your own e-Commerce site, you should consider provisioning your web analytics tool to generate one or more of the reports below:

  • Trends – compare annual, quarterly, monthly trends to determine if there are unexpected behaviors
  • Device comparisons – the influx of mobile and tablet shoppers often skews CAR metric because these visitors tend to shop on one device and then complete purchase on the desktop
  • Cross-comparison of CARs (conversion, revenue, items) – difference between different variants of CAR metric may reveal unique preferences of your visitors

Deep Dive

The next frontier in CAR metric evaluation is the provisioning of your web analytics tool for more complex analysis of the CAR data sets. At the risk of sounding too theoretical, here are suggestions about the types of reports that will help you get actionable about CAR:

Channel and Segment Analysis

Comparison of CAR metrics across different channels (e-mail, social, ppc, ….) or different types of visitors will help you determine the effectiveness of your demand generation efforts and it might help you uncover unique preferences of different types of visitors.

Visits to purchase

This report should help you determine a distribution of the number of visits before the purchase is made. It will paint the picture about your consideration path, helping you develop UX improvement or targeting strategies that enhance ongoing engagement or buying urgency. Ultimately this data should help you discriminate between real buyers and ‘window’ shoppers.

Time to purchase

Similar to the visits to purchase reports, you can use time as another criteria to learn about decision-making properties of your visitors.

Funnel Modeling

The goal of this report is to leverage statistical regression techniques to identify kinks in the funnel, steps in the checkout funnel where the abandonment rate is higher then the minimum error line.

You cannot go wrong

Although the purpose of the blog was to help you evaluate your own CAR metric, it is hard to go wrong with the CAR optimization initiative. Any improvement in this area is twice as valuable as any improvement on the top of your e-commerce sales funnel.