guh.me - gustavo's personal blog

A summary of metrics for e-commerce applications

⸻

I “extracted” this article from a paper I had to write for my MBA, back in 2016. Please forgive me for any translation errors or if the article quality is subpar (I personally hate it, but it contains some good information, lol).

The advent of the personal computer and the internet has revolutionized the way people buy. The US Census shows e-commerce on the rise, representing almost 10% of US retail sales in the first four months of 2016, a phenomenon that is repeated in other parts of the world. However, these new digital sales channels also generate a large amount of data that, if not well used or misused, can significantly impact the performance of online businesses. The application of a diverse set of metrics can take advantage of this data, and can bring benefits such as greater agility in reaction to changes, forecasting trends, improving the consumer experience, maximizing profits and improving processes. Therefore, this article seeks to answer the following question: What are the most relevant metrics for e-commerce applications?

The sheer volume of data produced and captured by modern e-commerce applications hides a true treasure trove of information, and the 22 indicators set out in this article can help any e-commerce company to easily recover part of this treasure. This set of metrics is not complete - each distinct business and strategy needs specific metrics - but it provides a range of information that allows you to measure the performance of an e-commerce operation and support a successful business strategy.

To write this article, we did a bibliographic research that was carried out in technical-scientific books and in articles available in online electronic databases, such as Scielo and ACM, using the following keywords: business intelligence, ecommerce, analytics, key performance indicators.

Key Performance Indicators

Successful entrepeneurs can answer with confidence questions such as “How’s my business going?”, and they can back up their answers with data. In fact, it is only because of this ability that modern companies like Amazon can achieve success and remain competitive and growing steadily.

But what is the difference between a successful company and a mediocre company? The successful company knows what to measure and what to do with those measures. That is why each company must define its own performance indicators that align its operations with its strategical objectives - and that’s where the key performance indicators join the stage.

Key Performance Indicators (KPIs), are metrics that help understand a company’s performance. David Parmenter, a KPIs guru, says that “KPIs tell you what to do to increase the performance dramatically”, and that “KPIs represent a set of measurements focused into the most critical organizational performance aspects for the current and the future success of the organization”. He also establishes a few characteristics an indicator must have to be considered a KPI:

Parmenter also emphasizes that, when a monetary unit is used to measure something, it is converted into a result indicator. However, this is not an opinion shared by all authors, and for the sake of simplification, on this article we will consider measures with monetary units as KPIs.

Finally, we can also summarise a few steps to choose the ideal KPIs:

  1. Analyze the company corporate strategy: before choosing a KPI, learn what the company’s goals are, and how they will be measured in e-commerce.
  2. Analyze how e-commerce initiatives can help you achieve corporate goals and how they match the corporate strategy: how the KPI will assert that the corporate goals are being reached.
  3. Measure from the beginning and often: after defining what is success and the KPIs that measure it, start measuring to see if they provide the analytics value you’re looking for.

Metrics

Metrics can be separated in three groups:

Conversion Rate (CR)

The Conversion Rate (abbreviated as CR) is the most basic metric for virtual commerce. It shows how many visitors are converted into customers. To determine the Conversion Rate, the following formula is used:

$$CR = \frac{\#\ of\ conversions}{\#\ of\ visitors}$$

This metric can - and should - be segmented in several ways to display a better picture of the business. E.g.:

One of the most common problems of online stores is having a lot of traffic and few sales. The reasons for this vary: the site design does not pass trust, there is no “social proof”, high prices for products that look generic, traffic from inappropriate sources, lack of payment methods, etc.

Cost of Acquiring a Customer (CAC)

The Cost of Acquiring a Customer (CAC) represents how much money you need to spend in order to win over a customer. The lower the CAC, the better, since it is cheaper to conquer a customer. There are many different customer acquisition techniques like SEO, sponsored ads and social media, and they all cost either money or time.

$$CAC = \frac{Acquisition\ cost}{\#\ of\ new\ customers\ in\ period\ T}$$

Shopping Cart Abandonment

Shopping Cart Abandonment, measures the percentage of visitors who added a product to the store’s shopping cart but did not checkout. The lower the abandonment rate, the better.

The abandonment of cart can happen because of multiple factors:

This metric can be calculated with the following formula:

$$SCA = \frac{\#\ of\ people\ who\ added\ to\ cart}{\#\ of\ people\ who\ checked-out}$$

Average Order Value (AOV)

Average Order Value, also known as Average Ticket, represents the average amount of sales by a given customer or customer group during a period of time T. The bigger the AOV, the better.

$$AOV = \frac{Sum\ of\ order\ totals}{\#\ of\ orders}$$

AOV monitoring allows you to analyze how much revenue you can generate with the current traffic and conversion rate. This allows the creation of strategies to stimulate the increase of the average value of orders, such as bundling, cross-selling and up-selling.

Gross Margin

Gross Margin is nothing more than store revenue subtracted from the cost of goods sold (COGS), and is represented as a percentage of the product sale price - more specifically, it is the portion of the sale price that represents the store profit.

$$Gross\ margin = \frac{Revenue}{Cost\ of\ goods\ sold}$$

The literature shows that it is of the utmost importance to maintain a gross margin greater than the CAC for the business to be sustainable; also profit margin results can help build the store mix of products: there must be a combination of products that bring traffic to the website, but with lower profit margins, and high-performance products that are much more profitable even if they sell little. Also, you can try bundling to balance the profit margin between different products.

Customer Lifetime Value (CLV or CLTV)

Customer Lifetime Value (abbreviated as CLV or CLTV) measures how much a particular customer will spend in the store during their “life cycle”. Not all customers are the same, and that’s why it is necessary to focus on the customers who bring more value to the store.

$$CLV = \frac{Revenue\ per\ customer}{Customer\ acquisition\ cost}$$

The CLV calculation allows you to perform certain actions focusing on the customers that have the most potential to increase the store revenue.

Retention Rate

The Retention Rate allows you to monitor the company’s efficiency in attracting and retaining customers. Existing customers are much more profitable than new customers as long as they are satisfied with the store - which is essential to maintain a high retention rate.

$$Retention\ Rate = \frac{Customers\ acquired\ in\ period\ T}{Customers\ who\ purchased\ again\ after\ T}$$

Refund Rate or Return Rate

The Return Rate is calculated as the percentage of products returned in relation to the total number of sales. Returns are a normal occurrence, be them for product defects and customer dissatisfaction or regret, but they do hurt the store profit margins, and can also hurt customer retention (because of low satisfaction).

$$Return\ rate = \frac{\#\ of\ returned\ orders}{\#\ of\ orders}$$

The segmentation of the Return Rate metrics can reveal low performing items that should be removed from the catalog. Also, defining a rigid return policy can help minimize this rate.

Support Rate

The Support Rate measures the number of visitors and customers who must interact with a support agent before making a purchase. If this indicator is too high, you may need to improve product information, store policies, shipping terms, and so on. Moreover, it is essential that the customer support channels are friendly and visible. (METRIL, 2016)

$$Support\ rate = \frac{\#\ of\ support\ requests}{\#\ of\ orders}$$

Best Products and Categories

The combination of various metrics (Return Rate, Conversion Rate, Gross Margin, etc.) allows some products and categories of the store to be evaluated holistically. Give that, you can take actions to improve the performance of these: highlight top-selling products, offer bundles of products that are usually bought together, create promotional actions to clean the inventory of products that do not sell, among others.

Revenue

Revenue, by itself, does not have possess good analytical value, since it only indicates that the company is selling, which is quite vague. However, when used combined/segmented with other data, it can help identify the best and worst sources of revenue in the store. Among the possible forms of segmentation, there are:

Number of Orders

The Number of Orders represents the number of orders made in a given period of time T. This number by itself has little value, but when segmented by date (day, week, month, quarter, year) or population, it can reveal customer purchasing patterns, assist in measuring the outcome of marketing actions and provide an overview of the overall “health” of the business.

Number of New Customers

The number of new customers comprises the number of customers who have registered in the store in a period of time T. A “healthy” business will have a rising number of new customers (to compensate the customer churn), while a declining business will have a decreasing customer growth. Associated with other metrics, such as customer churn and customer lifetime value, this metric can reveal problems with customers acquisition.

Customer Churn

Customer churn (also know as churn only), represents the percentage of customers who buy once and never buy again in a period of time T, and represents the opposite of the retention rate. Customer loyalty guarantees the sustainability of a business, so the lower the turnover, the better. A low turnover is directly associated with a high CLV.

$$Customer\ churn = \frac{\#\ of\ customers\ who\ bought\ only\ once}{\#\ of\ customers}$$

Customer Retention Rate

This metric, basically, calculates the number of active customers in your business. This is easier to calculate if your business is subscription-based, but if it’s not, you can try a myriad of formulas to see which one fits your business better. Essentially, this metric can help you assert if your marketing and customer care efforts are fortifying your business or just wasting money.

$$Retention\ rate = \frac{\#\ of\ customers\ who\ bought\ more\ than\ once}{\#\ of\ customers}$$

Number of Visitors

The number of visitors is usually correlated with the number of new customers and with revenue. By itself, this metric has little analytical value, therefore, it must be used in combination wiht other metrics (e.g. revenue). In a “healthy” business, the number of visitors should be rising or, at least, constant.

Order Gap Analysis (OGA)

Order Gap Analysis, abbreviated as OGA, is the average time that repeat customers take to make a new purchase. This metric enables you to determine your customers’ buying patterns and, consequently, refine the store’s marketing strategies. Furthermore, the OGA allows the store to find out which products to focus on to grow in the long run.

Days and Visits Before Converting

This metric allows you to assess the actual behavior of customers in the store, and how long they take, on average, to place an order. This permits us to improve the marketing strategy of the site, which allows us to apply tactics to accelerate this process, such as those that approach the possibility of a product going out of stock soon or increasing in price in a near future - strategies that leverage the so called fear of missing out (FOMO), in order to make the customers more likely to take action.

Support Time

Support time refers to the average time that support calls take until the customer problem is resolved. Complementary to the support rate metric, this metric eases the identification of bottlenecks in the support processes and finding operations that could possibily be automated.

Visitor Referrals

Visitor Referrals are the store’s traffic origins and literally show how the customers reached the website. This information allows us to find out what marketing strategies work better and bring more customers, and also can help the development of more efficient and engaging customer acquisition strategies.

Exit Pages

Exit pages are the last pages that customers visit before leaving the store. Knowing when and where customers leave the store allows you to analyze the sales funnel and find out which parts of it may need improvement in order to keep customers’ attention. Such improvements may include adding more relevant information, removing irrelevant information, removing unnecessary steps, and reviewing company policies.

Bounce Rate

The bounce rate measures the percentage of users who left the store immediately, possibly because they did not find what they were looking for or because the store is too complicated to use. The bounce rate, in association with metrics such as referrals, allows the identification of inappropriate acquisition channels and provide feedback to adjust customer acquisition marketing actions. The bounce rate can be calculated as follows:

$$Bounce\ rate = \frac{\#\ of\ visitors\ who\ leave\ immediately}{\#\ of\ visitors}$$

Payment Methods

The data on payment methods offered by the store can be used to create many advanced metrics, such as:

References