About Funnels

 

Introduction

The purpose of this document is to impart additional information related to the concept of funnels to better empower you to leverage funnels in FullStory and beyond. While this information applies to funnels generally we hope you will leverage this knowledge to get more value from FullStory’s Conversions product as well as FullStory Segments. It is not a requirement that you be familiar with FullStory’s Conversions and Segment functionality in order to understand this document. (You can learn the basics about Conversions in this series.)

In this document we will discuss:

• What is a Funnel?
• Selecting what Funnels to build
• Beginning with the End
• Funnel Observation (using session replay)
• Working Backwards
• URL Funnels vs Event Funnels vs “Hybrid” Funnels

Let’s get started!

What is a Funnel?

Before we talk in more depth about funnels, let’s take just a moment to clarify what we mean by “funnel.” Funnels (or often “conversion funnels”) can be defined in a number of ways that are often industry specific. Here at FullStory we think of funnels in the broadest sense possible.

A funnel is simply a series of actions you want an end-user to take or a series of locations (URLs) in your digital property you want end-users to visit.

Navigate to Funnels to start building.

Below is an example of a basic checkout funnel visualization built in FullStory:

Depending on your business, you may have many different funnels on your site or app related to important end-user behaviors. One type of funnel might follow an end-user as they become a paying SaaS customer or sign up for a subscription of some kind. Another type of funnel may follow a new customer through their initial onboarding and setup process. The key takeaway here should be that funnels are meant to be specific to your business and should reflect meaningful moments in the relationship between your end-users and your business.

What Funnels to Build

To “build” a funnel means to define the necessary steps involved in reaching the desired goal.

Given that a funnel can actually include any series of actions or site/app locations, it may be hard to determine what funnels to start with when it comes to analysis and optimization. In order to pinpoint meaningful funnels in your digital property, simply ask yourself: What do I want people in my product (or on my site) to accomplish?

If you work in SaaS or have a subscription based business model then maybe an answer to this question is something like “sign up for a free trial.” If you are in eCommerce, maybe the answer is as simple as “buy something.” However many ideas you come up with, you will want to select one of those as the candidate for your first funnel. Since our demo site is an eCommerce business called The Fruit Shoppe, we will use the example of “buy something” for the purposes of this tutorial.

When building out a funnel for the first time it can be helpful and clarifying to work backwards and to limit the funnel to key events only. Excluding any optional steps that may occur along the way can help with clarity. So, let’s begin at the end!

Beginning with the End

Every funnel will have an endpoint, and this is often the best place to start when thinking about a funnel for the first time. For this example we want to think about the last thing to happen when an end-user buys something. Every eCommerce site is different, but this may be a final URL or an activity taken by the end-user (we will talk more about URL funnels vs activity funnels later on).

In the case of our example site (The Fruit Shoppe) we are going to use our order confirmation page as our funnel endpoint. When an end-user in our store clicks “purchase” then an order has been triggered and they should be sent to our confirmation page (url path /confirm). These sorts of success pages make great funnel endpoints.

Example of building a funnel, ending with our confirmation page (url path /confirm):

Funnel Observation with Journeys and Session Replay (Optional)

If you are unsure what important actions precede your funnel end-point, it is a really good idea to use Journeys or session replay to learn more about common navigation paths on your site.

Journeys help you visualize how users navigate. You can create Journeys that show you the most common paths users take to reach a given end point (or, the most common paths they take from a given starting point) to identify steps to include in your funnel. 

Session Replay shows you a video-like reproduction of actual user sessions on your site. By watching several sessions, you can also also gain insight into the steps users commonly take. You don’t need to include every single thing a user might do when building out your funnel, but this research will help to inform exactly what key steps you include. For example, if there are fields in a form that need to be completed between clicking “checkout” and clicking “purchase,” you don’t need to include each of these “micro” actions in the funnel because the "macro" action of completing and submitting the form necessarily includes all the individual form fields.

If you already know exactly what you want to include in your funnel you can skip these steps. However, many people learn a lot from these exercises. It is not uncommon to learn, for example, that there are actually two or three different key funnels (or paths) that terminate at a common final step. (Guest checkout vs Member checkout, for example.) If something like this is the case for your digital property, then you will actually need to build and analyze several funnels to get a more complete picture of your end-users’ common journeys.

Working Backwards

Now that you have a solid funnel endpoint, you can work backwards from there to fill in the other key navigation steps in your funnel. In our Fruit Shoppe example, the key step before clicking “purchase” is clicking the “checkout” button. The key action that precedes clicking “purchase” is clicking “Checkout,” and so on. Remember, there are probably things a user will do between each of these key steps (like filling out the Name and Shipping Address fields) but these actions can be excluded from the funnel.

End at the Beginning

Continue to work backwards in this way and you will eventually come to a place you can consider the beginning of your funnel. But, how do you know where to stop? There is no hard and fast rule for this and it can depend a lot on how your site is built. Some people like to begin their funnel at the home page (or wherever people land on your site initially), while others like to choose the first end-user action or URL that more strongly indicates intent to convert (like adding a product to cart).

In our Fruit Shoppe example, end-users who buy something must add products to their cart from the /market page. Because this is the case, we have decided to use visiting the /market page as the first step in our “buy something” funnel.

By starting at the funnel endpoint of visiting the order confirmation page (URL path /confirm) and working backwards, we created the following funnel for end-users who buy something on The Fruit Shoppe:

Example of building a funnel, beginning with our market page (url path /market):

URL Funnels vs Event Funnels

In our example funnel above you may have noticed that we included both URL visit information and end-user click event information. When learning to think about funnels for the first time, it is common for people to think in terms of either URL information or in terms of end-user event information like “clicks.” Rarely do people intuit that these two different types of information can be used together in a single funnel. In this section we will explore what it looks like to use each of these types of information alone and will then explore some of the benefits of using them together.

URL Funnels

For many people this is the default way to think about constructing funnels. You can think of a “URL funnel” as a funnel where each step is defined simply as having visited a given URL or URL path.

Event Funnels

While URL funnels can be interesting and useful in certain situations, they are not the only way to go about analyzing how end-users are navigating your product. Once you have digital experience analytics (like those available from FullStory) it becomes easy to start thinking about funnels in terms of end-user actions. You can think of these “event funnels” as a series of actions taken by the end-user (often clicking something).

Telling Different Stories

How you decide to build the funnels can have downstream consequences for how you understand your product. To help us explore this concept, let’s compare our two example funnels from above side by side. 

Visited URL Funnel

Clicked Text Funnel

 

What you might notice about these two different funnels is that they are both built to look at the same end-user journey through our eCommerce site The Fruit Shoppe. This journey begins on the home page of the site and ends with a product being purchased. Both of these examples are completely valid approaches for building a funnel and looking for points of friction in the end-users’ journey. However, they will not show you exactly the same information. Let’s add to our funnel examples the number of end-users who completed each step and look at them side by side again:

Visited URL is www.fruitshoppe.com	5.7k	Clicked text is “Market”	5.2k
Visited URL path is /market	4.7k	Clicked text is “Add to cart”	5.2k
Visited URL path is /cart	3.9k	Clicked text is “My cart”	4.4k
Visited URL path is /checkout	3.1k	Clicked text is “Checkout”	3.7k
Visited URL path is /confirm	1.1k	Clicked text is “Purchase”	2.3k

 

When we start comparing the numbers of end-users at each step of these funnels, it immediately becomes clear that these funnels, while largely “the same,” are telling us very different stories about the same group of people. Let’s dig into one specific way these two funnels differ.

Checkout -> Purchase Conversion

The difference between our two funnels at the last step is both startling and important.

Visited URL path is /checkout	3.1k	Clicked text is “Checkout”	3.7k
Visited URL path is /confirm	1.1k	Clicked text is “Purchase”	2.3k

 

Our URL funnel is telling us that only 35% of end-users that make it to the checkout page end up actually purchasing something. For most businesses it would be highly problematic to see a drop off at the last step in a purchasing funnel. This kind of loss at a single step is a red flag and would make the /checkout page a prime candidate for immediate improvement.

What is interesting is that when we look at our event based funnel, things look quite different. Our event based funnel is telling us that 62% of people who make it to the checkout page actually follow through with purchasing something. This is a much more acceptable conversion rate than the 35% shown in the URL funnel, and while there is still room for improvement, a finding like this is much less likely to cause a red alert.

This comparison is a great example of exactly how the funnel you choose to build can affect everything from the high level understanding you have of your product, to how and where you decide to spend time and resources optimizing your product.

"Hybrid" Funnels - Using URLs and Events Together

As illustrated above, how you decided to build your funnels can have serious implications for your business. From understanding how your products are performing to where you spend your time. As a result it is important to get as much clarity as possible. Enter “hybrid” funnels.

In our previous examples we looked at both a URL funnel and an Event funnel of a single journey through our site. Both types of funnels tell a different story about what is happening on our eCommerce site Fruit Shoppe and one story is much more worrisome than the other. A “hybrid” funnel actually takes both types of information and combines them into a single, more granular funnel. Using this approach allows you to get a more accurate understanding of what is occurring in your product than using either URLs or events alone. And, you have the ability to uncover issues you would have never found using one of those other funnels. Our “hybrid” funnel, built using the information from above, looks like this:

Visited URL is www.fruitshoppe.com	5.7k
Clicked text is “Market”	5.2k
Visited URL path is /market	4.7k
Clicked text is “Add to cart”	5.2k
Clicked text is “My cart”	4.4k
Visited URL path is /cart	3.9k
Clicked text is “Checkout”	3.7k
Visited URL path is /checkout	3.1k
Clicked text is “Purchase”	2.3k
Visited URL path is /confirm	1.1k

How awesome is this?! A single funnel with double the information. Not only is it more accurate, but we can learn things about our product we would have totally missed before. Let’s explore a couple examples.

Both Ends of Navigations

In products (like our Fruit Shoppe site) end-user navigations have a start and a destination. By this we mean clicking the “Market” button takes you to the URL path /market. When we were using our shorter separate funnels previously we were only looking at the start or the end but not both. In our hybrid funnel we get to see both ends of each navigation. We have highlighted two instances of this here in red:

Ideally, when end-users click a button in your product or site it should always take them to the desired destination. However, in the example about, when people click “Market," something goes wrong. 5.2k people click “Market” but only 4.7k actually land where they should on the /market URL path. This is a sign that something is going wrong and needs to be investigated. This is the kind of issue that would go completely unnoticed in our URL only or event only funnels from before.

This kind of issue is also what is responsible for our very different stories around end-users completing purchases. For some reason, users that clicked “Purchase” were not all landing on our confirmation page. This lead to very different looking conversion rates at the last step of our earlier, shorter funnels.

Additional Non-critical "Micro" Steps

Another great thing about “hybrid” funnels is they allow you to easily layer in lots of details that are not critical to funnel completion.

An example of this in our funnel are the clicks on the “Add to cart” button:

In our hybrid funnel, this step is not necessary for understanding how people navigate the purchase funnel as a whole. We could exclude this step and still be able to get a decent understanding of how the overall funnel is performing, but the additional insight you get from including this step can be helpful. In this case, 17% of end-users who make it to the /market page do not add something to their cart. It might be interesting to explore different ways to improve this metric.

After seeing this example you can start to think about what other optional steps it might make sense to include in your funnel. For eCommerce, these might be watching a product description video or interacting with the product review widget. Both of these behaviors could potentially impact conversion positively or negatively but are not required navigation events in the purchase funnel.

Final Thoughts

Hopefully this exploration of the concept of funnels has left you feeling empowered and ready to tackle your product’s user journeys with a fresh perspective. If you are new to thinking about funnels, remember to start simple and build. Begin at the end!