Summary: In the first part of this series, we introduced analytics, metrics, and KPIs. We discussed why metrics are important, how to interpret them, and how to decide which ones to track. In this article, we’ll review the metrics that we have found to be most useful to our retail customers. We’ll discuss how they are generally calculated, how they can be improved, and when possible, any industry benchmarks that are worth considering. We’ll also touch on the use of segmented analysis to provide a clearer picture of business health. We’d encourage you to start with part one, An Introduction to E-Commerce Analytics, if you are new to analytics. If you’re already familiar, then part two is a great starting point. In part three, we’ll dive even deeper and discuss some technical tips for collecting, analyzing, and acting on your metrics at scale.
There are metrics that can provide you performance information from every operational group within your business and from every end of the supply chain. We may tackle metrics related to manufacturing, logistics, human resources, or finance in future posts, but for now we’ll focus specifically on e-commerce storefront performance, which is almost entirely built upon sales, marketing and customer experience. Of course, all these departments are interconnected, so some metrics we’ll review require information from multiple areas, but the general focus is the same — buying and selling products, and interacting with your customers.
Below are the metrics we’ve found to be most useful when assessing the health of your e-commerce storefront. Before running off and collecting these, be sure to review the next section on segmentation, because how you view the data is just as important as the data itself!
Arguably some of the most important metrics in e-commerce, conversion rates measure the fraction of users that move on to the next step in the buying process from some previous step. These can be tracked all throughout your sales and marketing funnel and are most commonly reported as a percentage. Some common measurement points include:
Conversion rates are easy to track with modern tools, and they are highly actionable, but each one requires its own troubleshooting, so careful review of the data is critical. We could write an entire series on conversion rate optimization (CRO), so we won’t dive too deep here, but know there are countless free resources available online, as well as entire products and companies that have been built for increasing conversion rates. For quick benchmarking, it’s helpful to know that e-commerce stores tend to see an average bounce rate of 45.7%, an average abandoned cart rate of around 70%, and an average overall conversion rate of 2.27%.
One metric that massively impacts customer experience, and ultimately conversions, is page load time. In fact, within the first 5 seconds of attempting to access a website, conversion rates drop by 4.42% per second of additional load time. This is particularly scary when mobile websites, according to Google, have an average load time of 15.3 seconds! This critical metric is not only key to your business’s success, but it’s also extremely easy to measure, making it one of our top choices for metrics to pay attention to.
If you really want to cut your site’s load time, consider switching to a headless architecture. This allows you to use best in class technologies and optimize your site for whatever device, screen size, or connection speed lands on it. Furthermore, since your front end isn’t reliant on the same infrastructure as your back end, heavy front end loads - say during product drops - are less likely to impact the infrastructure that manages orders and payment. Making the switch to headless can be quite an investment though, so if you’re looking for some smaller steps, two easy improvements are removing unused plugins from your site and reducing media file size where you can.
Like the name implies, the customer acquisition cost (CAC) is the amount of money it takes to acquire a new customer. Knowing how much you’re spending to make a sale is critical in determining whether that sale is actually worth making. However, we’re rarely ever interested in a single transaction. Instead, the goal is to cultivate a relationship with the customer that leads to many future purchases which justify the initial acquisition cost. For this reason, the ratio of CLV (customer lifetime value) to CAC is often observed, and it is generally thought that a minimum ratio of 3:1 is a fair target, though of course it depends on the situation. Early phase companies may tolerate a lower CLV/CAC ratio, but caution is necessary here – the CAC is often called the “startup killer”, because it can destroy margins at an alarming rate.
To measure CAC, the most common approach is to divide your total marketing spend by the amount of new customers you’ve obtained during the same period. To get more granular, you can use tools that track attribution of each conversion back to a source, though this can get tricky when you consider that conversions often involve multiple touch points across channels. Moreover, some individual channels such as print ads or experiential marketing, are particularly challenging to track. There are some clever approaches to this, however, which we’ll touch on in part three.
If you want to improve your CAC, effective targeted marketing is key. Find channels and messages that speak to your audience, or combinations that speak to each audience segment separately, in order to increase conversion rates and drive down your CAC. Some companies take a bottom-up approach, setting a maximum CAC they will tolerate in order to maintain a certain margin and then assigning a marketing budget based on that. This of course requires some level of confidence in the conversions you are going to get out of the spend, but over time you will indeed be able to make such predictions based on past performance.
Okay, so you’ve successfully convinced a customer to buy from you. Although we aspire to long-term relationships, on average only a third of customers will make a repeat purchase within their first year, so it’s important to make every purchase, including that first one, really count. This is where average order value (AOV, sometimes called average order volume) comes in. The goal is to maximize the amount spent per order, thus making every related variable cost less impactful on your bottom line. Measuring AOV is simple – just divide your total revenue by your total number of orders within the time period of interest. Generally the higher your AOV, the healthier your online store (assuming your product is profitable).
Ways to improve AOV are:
E-commerce industry benchmarks for AOV tend to land between $50 and $200 depending on the industry. Combined, reports from across the web indicate an industry average AOV of $128 in 2020.
A perfect customer would buy from your brand consistently for the rest of their life. Of course this customer does not exist, but we can measure how close we are to this model by monitoring customer retention rate (CRR), generally defined as the fraction of customers who remain loyal after a purchase. Retention rate is easiest to measure for subscription-based sellers who determine purchase frequency on their customers’ behalf (usually providing monthly or annual subscriptions). In this case, retention rate is simply the percentage of customers who continued their subscription the following period. For more general online shopping, however, the customer determines purchase frequency, so calculating CRR can be more complicated depending on how accurate you want to get. Some companies in this setting report retention rate as the percentage of customers who have made at least two purchases within a one-year period. Some just look at all customers over time and see which fraction ever made a repeat purchase. Sometimes it is valuable to calculate retention rates based on how many total repeat purchases have been made, as the likelihood of an additional purchase increases with each purchase made from a single customer.
Ultimately, we suggest tracking retention rate for the purpose of predicting CLV, which we’ll get to next. For this, the time period in which retention was measured is actually meaningful, so we recommend against measuring over all time. The exact time period should be dependent on your business and what is meaningful for your industry. For example, tracking repeat purchases within a year might be unrealistic if you’re selling cars or houses but might make perfect sense for household goods or cosmetics. While it may be more accurate to track retention based on each additional purchase made, this requires a lot of data to get statistically significant values, so in most cases it’s best to stick to the simple binary approach – did the customer return after their first purchase? Yes or no.
Obviously, the higher your retention rate, the better off you are. Retaining customers is immensely less expensive than attracting new ones, and by retaining them, the impact of that initial CAC is diminished over time. To increase retention, post-purchase touch points are key. Mailing list follow-ups, loyalty programs, and exclusive offers for your existing customer base are great ways to drive retention. Due to high competition, e-commerce retention rates tend to land around 30%, much lower than the 80%+ rates seen in other industries such as insurance or media, but some breakaway e-commerce companies can achieve very high retention. As an extreme example, Amazon retains 90% of its customers year over year.
Relevant metrics vary from business to business, but in almost every case, customer lifetime value (CLV) is very high on the list. CLV is the amount of value gained from a single customer for the entirety of your relationship. It is commonly used to project future earnings based on your customer base and to help justify ongoing acquisition and retention costs.
Some companies with significant historical data will retrospectively measure CLV by dividing their total revenue by the number of customers involved in those purchases. This provides limited insight, as a long history includes data from when conditions such as customer behavior, market forces, even your own products and prices were different. On the other hand, a short history provides even less value, as long term behavior is entirely unknown.
An alternative method, and the one we’d recommend, is to use data from a representative near-term time period and combine it with metrics discussed above to predict CLV over the next few years. You’ll first have to calculate the expected customer lifetime in units of time by subtracting the retention rate from 1 and then dividing 1 by the resulting value. (See formula below. Note that the units will be dependent on your CRR time period.)
Once you have the expected lifetime, you simply multiply that by the average order value and by the average purchase frequency (defined as the number of purchases a customer makes on average in the measured time period) and you wind up with your customer lifetime value in units of currency.
Side note: This formula is a good starting point and thus is our recommendation to the general e-commerce business owner, but it does have a few holes. For one, it assumes that both retention rate and purchase frequency will remain constant as years go by, and we know that at least the first of these assumptions is false. It also does not account for the time value of money. If we are discussing particularly long customer lifetimes, then the future revenue expected should be discounted to present value for a more realistic comparison against present day costs. Lastly, it only considers top line value. It can often be useful to consider the actual contribution margins of the purchases and to subtract out some risk values for expected returns or refunds. All of these considerations can be accounted for by varying the general CLV formula, and the best method, as usual, will depend on your specific business needs and the time it’s worth for you to invest in such calculations.
Ultimately, there are three ways to increase customer lifetime value, each usually achieved through marketing and incentivization:
Industry-wide benchmarks for CLV are hard to come by and of minimal value, which is why it’s most common to track CLV as a ratio to your CAC. That said, at least one report has indicated an average CLV of $168 across direct-to-consumer e-commerce brands.
Return rate, or refund rate, is a critical metric to track because a high value can be detrimental to your business in more than one way. First, it means you are losing sales. Second, you are adding costs by facilitating these returns. And third, it might mean you are generating unhappy customers who will talk negatively about your brand. Don’t expect to drive this value to zero though, as returns have become a natural part of the shopping experience. 79% of customers expect free returns, and 67% of shoppers check the returns policy before making a purchase from a new seller. In e-commerce, you can expect an average return rate between 20-30% depending on the season, compared to 10% or lower in a brick and mortar store. The easiest way to drive down return rate is to limit surprises by providing ample information about the products up front. If a return is requested, try offering an exchange or store credit at least once for revenue retention. Whatever your return policy is, tracking your return rate and capturing the reasons for returns is crucial in order to monitor the extent of their impact and determine how best to limit them.
While we’re mostly skipping operations in this post, one operational metric that is closely tied to your storefront sales and is therefore worth an honorable mention is sell through rate (STR). STR is the percentage of inventory that you move through within a period of time. The amount of time you choose to measure will dramatically impact the STR value, because you sell more inventory as time passes. For example, one report suggests an apparel industry average sell through rate of 24% over 8 weeks, 29% over 13 weeks, 46% over 26 weeks, and 67% over a year. This means that it will be important to stay consistent in the values you use to compare over time or between products, departments, and companies. Other inventory-tracking metrics to consider are inventory days on hand and inventory turnover rate. All of these track, in one way or another, the time it takes to sell through your stock, but the general goal is the same – strike a balance between selling products quickly enough to avoid discounting them and restocking regularly enough to avoid the dreaded “out of stock” label.
Our last key metric to make the list is the infamous NPS, or net promoter score. Created by Bain & Company in 2003, the NPS quickly climbed the ranks of customer satisfaction measurement methods and is still near, if not at, the top today. The beauty here lies in its simplicity. Ask your customers to respond on a scale of 1-10 how likely they are to recommend the purchased product or service to a friend or colleague and place them into one of three buckets based on their response:
Your NPS is then the percentage of customers who are promoters minus the percentage of detractors, hence the “net” promoter score. This method has proven to be both a minimally invasive survey method and a very good indicator of actual customer satisfaction. As of 2021, the industry average NPS for e-commerce was 45. A value above 30 is traditionally considered to be satisfactory, and above 70 is considered excellent. As always though, tracking against your own improvement and against benchmarks that make sense for your specific business is key. However, if you have a negative NPS, that is you have more detractors than promoters, this is a bad sign no matter who you are, and some changes to your product quality or customer service are necessary in order to increase customer satisfaction.
Before we wrap up, it’s important to mention one critical analysis method you’ll need in your toolkit – segmentation. A company-wide or catalog-wide metric run on your full data set is hardly ever the most useful version. Instead, consider stratifying your data into segments and comparing between them, the most common segments being attribution (source/marketing channel), customer location or demographics, on-site behavior, and product or product category. For example, you may find that NPS is higher for your west coast customers than east coast, indicating that your products are better suited for them or perhaps they are having a more relevant shopping experience. You may find that your acquisition costs are significantly higher for customers who convert from Instagram ads rather than your referral program, or perhaps customers who land on your about page tend to spend more (higher AOV) than those who don’t. All of these insights will help you understand your customer and your business better, and you’ll be able to use them to optimize detailed aspects of your e-commerce storefront.
We’ll get more into the methods for collecting and analyzing these data at scale in part three of the series, but for now you may want to consider setting yourself up for successful collection and segmentation by owning your 1st party data, enriching that data for more valuable insights, and investing in a customer data platform (CDP) to manage your data. These tools are becoming increasingly important as we move away from the era of 3rd-party data services, and they’ll provide you with the flexibility to dig into your data and segment it any way you need in order to gather the information that will help you grow your business and turn your customers into brand evangelists.
This concludes the second part of our 3-part series on e-commerce analytics. By now you should have a good sense of why metrics are important for your business to track, how to interpret them, and which metrics might be useful for monitoring the health of your e-commerce storefront. In the third and final article, we’ll provide some practical tips for how to set up your analytics pipeline, covering tools and techniques you can use to collect and analyze your data at scale, automate critical process steps, share results with your team, and take action on your findings. If you’re ready to get started on your analytics journey, we are always here to help!