Introduction:

Are you facing challenges in maximizing your e-commerce conversion rates? You’re not alone. Conversion Rates (CR) are the cornerstone of success in the digital marketplace. This article is your go-to guide for understanding and leveraging essential metrics and Key Performance Indicators (KPIs) to elevate your e-commerce performance. We’ll explore strategies to decode the dynamics of CR and optimize them for tangible business growth.

When it comes to developing products, having the right metrics and key performance indicators (KPIs) is essential for assessing your situation, identifying strengths and weaknesses, and pinpointing areas for improvement. This systematic approach empowers you to make data-driven decisions, elevating the quality of your choices and facilitating faster and more precise growth. In the realm of e-commerce, one of the pivotal metrics that demands your attention is the Conversion Rate, often abbreviated as CR. The term “Conversion Rate” signifies a ratio obtained by dividing the number of desired actions, such as making a purchase, adding a product to the cart, or clicking on a video, by the total number of customers or visitors who engage with your platform or website.

Defining the appropriate conversion rate or other relevant metrics represents the initial step in your journey. However, the next crucial phase is analysis once you’ve defined and measured the conversion rate. Understanding why the conversion rate has either increased or decreased over the past days, weeks, or months becomes imperative. This step is often pivotal in understanding the situation, significantly when your website has scaled up in various dimensions.

For instance, if you operate across multiple platforms such as desktop, mobile web, and applications, offer diverse product categories, or provide various customer journeys leading to your products, the complexity of analyzing the data intensifies. As your product scales, the need to break down your primary objective into more discernible terms becomes even more crucial. In this blog post, I will elucidate how we harnessed mathematics to deconstruct our primary objective into interpretable components, enhancing our analytical capabilities.

Define objectives and scenarios: 

Imagine running an online store that sells a diverse range of products, from cutting-edge electronics to fashionable apparel. The key question is: how do you measure success and growth? Let’s define two practical conversion rates – ‘Order to Session Rate’ and ‘Items to Page View Rate’ – to understand how your website transforms casual browsing into successful sales. By setting these benchmarks, we can navigate through the complexities of e-commerce, ensuring a focused approach to boost your sales numbers.

Order to Session Conversion Rate: 

This metric aims to determine the daily number of orders generated per unique session. It assesses the efficiency with which each session translates into an actual order.

Items to Page View Conversion Rate: 

This rate measures the daily quantity of items sold per page view. It’s a metric to understand the effectiveness of each page view in encouraging an item’s purchase.

It’s important to note that both metrics are computed by aggregating user behavior across the board. The numerator (orders or items sold) represents the total sum, whether it stems from a single user or a multitude.

Aspects for Analysis:

Different platforms and product categories play a crucial role in shaping your conversion rates. For instance, the way customers interact with your website on a desktop might differ significantly from their behavior on a mobile app. Similarly, the conversion rate for electronics might vary from that of apparel. By dissecting these metrics, you can gain valuable insights into optimizing your e-commerce strategy, ensuring that every visitor’s journey on your site has the potential to culminate in a sale

Breakdown CR:

Before we dissect our Conversion Rates (CRs), let’s reevaluate our formula to discern the interplay between our principal CRs and the various aspects we’re interested in. For example, we can examine the relationship between the first CR and different platforms:

Our initial CR formula is:

CR = Total Orders / Total Session

Breaking it down by platform, we get:

CR= {Total Orders of Application + Total Orders of Desktop + Total Orders of Mobile web} / {Total Session of Application + Total Session of Desktop + Total Session of Mobile web}

We can further refine this to:

​where 

• and  are the Orders and Sessions, respectively, for platform i.
• = is the CR for platform 
• =  represents the proportion of sessions that belong to each platform i

  Hence, indicates the session share belonging to each platform. Consequently, we can express the total CR as a weighted sum of the individual CRs for each platform:

  

This formulation allows us to view the total CR as a composite of each constituent platform’s CRs and session shares, providing a more nuanced understanding of the influences affecting the overall rate.

When analyzing the change in Conversion Rate (CR), we’re interested in understanding how different factors contribute to its variation. Breaking down the differential of CR into its components can help us do that. The differential of CR can be expressed as:

This equation represents the total change in CR as the sum of changes in the product of the share of sessions (wi) and the (CRi)platform-specific across all platforms. Each term within this summation has a distinct interpretation:

• 
  • This term represents the impact of the change in the share of sessions (wi) for a particular platform on the total CR. It shows how much the total CR is affected when there’s a shift in the distribution of sessions among platforms while the platform-specific CRs remain constant.
•  
  • This term indicates the impact of the change in the platform-specific CRi on the total CR, assuming the share of sessions wi remains constant. It reflects how variations in the efficiency of each platform (in converting sessions to orders) influence the overall CR.
• 
  • This term captures the concurrent changes and their effects on the total CR. It represents the interaction between the changing share of sessions and the changing platform-specific CRs. This term is essential for understanding the compound effect when both the distribution of sessions and the efficiency of conversion are changing simultaneously.

By breaking down the change in CR into these components, you can better understand what’s driving the changes in your total CR and better identify areas for improvement or investigation.

To visualize how changes in the components wi  (the share of sessions) and CRi (the platform-specific conversion rate) affect the total Conversion Rate (CR), imagine a rectangle where the length represents CR and the width represents w. The area of this rectangle (w⋅CR) represents the contribution of a particular platform to the total CR at time T As time progresses, both w and CR might change, leading to a new, larger rectangle. This change can be visualized as additional areas adjoining the original rectangle:

•  Red Area ( Δw  ⋅CR  ): This represents the change in the contribution to total CR due to a change in the share of sessions (w), while the platform-specific CR remains constant. The red area is the additional width added to the original rectangle, extending its base.
• Blue Area  ( w  ⋅ ΔCR  ): This represents the change in the contribution to total CR due to a change in the platform-specific CR, while the share of sessions (w) remains constant. The blue area is the additional height added to the original rectangle, extending its length.
• Yellow Area ( Δw  ⋅ ΔCR  ): This area represents the concurrent changes in both the share of sessions (w) and the platform-specific CR. It’s the extra corner piece that appears when both dimensions of the rectangle increase.

To further clarify the contributions of each platform to the change in total Conversion Rate (CR), the equation can be restructured. While the previous formulation helped us understand the different components of change, this revision allows us to focus specifically on how much each platform is responsible for the change in total CR. The revised equation is

By summing these terms for all platforms, the equation demonstrates the total change in CR while providing insight into the contribution of each platform to that change. This breakdown allows for a more nuanced analysis, helping identify which platforms are driving changes in the overall conversion rate and how.

The table below is an example of breaking down the change of CR between two arbitrary periods of time. As you can see the change comes from changing the CR of each platform not changing in share of view . Also, you can see each platform’s influence on changing the total CR. 

Platform	diff weight	diff CR	diff Weight & diff CR	Platform Coefficient	Platform influence	Diff Total CR
Android	4.2	107.5	0.5	112.2	57%	0.1963%
IOS	-0.7	1.7	0.0	0.9	0%
Mobile Web	-3.64	49.5	-0.5	45.40	23%
Desktop	0.9	36.8	0.2	37.9	19%
	0.7	195.5	0.2	196.3
Terms influence	0.03 %	99.5%	0.47%

 

Summable Precondition

Understanding the concept of a summable Conversion Rate is like solving a puzzle. Each element – be it a product category, customer segment, or journey phase – must integrate seamlessly to offer a complete picture of your e-commerce performance. This holistic view is essential for accurate analysis and strategic decision-making, helping you identify and capitalize on every opportunity to enhance your conversion rates.

For instance, when dealing with the ‘Order to Session’ CR, we look at the number of orders per unique session. If we try to break this down by categories or customer journeys, we encounter a problem. Unique sessions can’t be attributed individually to each category or journey because a single session might encompass multiple categories or journey steps. Therefore, it’s not straightforward to assign a portion of a unique session to one category or journey without overlapping or missing out on some parts.

Instead, to analyze the contribution of each category or journey, we would use the second type of CR, like ‘Items to Page View’ CR. This metric is more granular and can be attributed directly and individually to different categories or journeys. Each page view is specific to a particular item or category, making it easier to break down and analyze how different aspects contribute to the overall CR.

In summary, ensuring that your CR is summable in the context of your analyzing aspect is a critical prerequisite for breaking it down effectively. It allows for a more accurate and meaningful analysis of how different factors contribute to changes in the CR.

Interpretation 

After breaking down your Conversion Rates (CRs) into different aspects, your analytical work truly begins. Now, you have the tools to uncover the ‘why’ and ‘how’ behind changes in your CR. By adding more aspects based on your needs, you can deepen your analysis. For instance, you might consider marketing journeys alongside internal customer journeys to see how they impact CR. You can investigate which factors most significantly affect CR during weekdays or special campaigns, or understand how the share of sessions and the cannibalization of certain journeys or categories influence your main objective.

It’s crucial to remember that while these breakdowns can illuminate how different elements influence the CR, they don’t necessarily indicate causation. They are tools to help you identify areas that warrant further investigation, guiding you to potential root causes of changes.

The flexibility of this approach allows you to extend your analysis further. For instance, you can explore joint distributions to gain insights into how different aspects interact with each other. You might break down journeys into categories to discover which categories are more popular in specific journeys. While it’s possible to conduct this analysis without examining joint distributions, doing so can provide a finer granularity and a higher resolution of analysis.

Ultimately, the level of detail and the aspects you choose to analyze depend on your specific needs and goals. The framework is adaptable, allowing you to set the granularity and resolution based on what you seek to understand about your CR. This flexibility is a powerful feature, enabling you to tailor the analysis to best suit your business’s unique context and challenges.

Conclusion:

With the insights gained from this comprehensive exploration of Conversion Rates, you’re now equipped to take your e-commerce strategy to new heights. Remember, understanding the nuances behind CR is key to making informed, strategic decisions. Apply these learnings to your digital platform, and prepare to witness a significant transformation in your e-commerce performance. The world of online retail is dynamic, and with these tools at your disposal, you’re ready to thrive in it

Remember, these breakdowns illuminate pathways and correlations, but they are starting points for deeper investigation rather than definitive answers. The real power lies in your ability to adapt and extend this framework to meet your unique business needs, whether that means incorporating new aspects, exploring joint distributions, or adjusting the granularity of your analysis.