There are many valuable uses of this type of data stream, and I plan to discuss many of them in future posts. In this post, I will describe how this data can be used for “granular funnel analysis,” or what I like to call, Digital Conversion Funnel 2.0.
Why are Digital Conversion Funnels So Popular?
Before getting into Digital Conversion Funnel 2.0, let’s summarize why so many marketers and product managers set up, track and analyze digital conversion funnels in the first place.
The conventional thinking is that a conversion funnel is the set of steps through which a visitor to a website typically proceeds before converting. A conversion, of course, is an action that you want your visitors to take on your site, such as buying something, downloading something or signing up for something.
Taking a standard e-commerce site as an example, the steps in the funnel usually include arriving at the site, browsing products on the site, adding one or more products to the shopping cart and finally checking out. There may be other steps involved, but e-commerce marketers generally select a few key steps and define their conversion funnel accordingly.
Marketers spend a lot time defining and redefining their funnels (i.e., selecting the steps that they think represent a customer’s journey until the conversion point), tracking them and analyzing them. Their objective is to identify “drop-off” points in their preconceived customer conversion journeys and to optimize each step of the way to maximize the percentage of visitors who convert. The idea is that every successful tweak higher in the funnel has big impact on the next stages in the funnel, because it means that there are more potential customers at the lower stage.
Using this approach to optimize websites is very popular primarily because it pinpoints specific parts of website on which a marketer or product manager can focus and take action: once funnel tracking is up and running, it becomes pretty clear where drop-offs are occurring and, thus, it becomes clear where to focus experimentation and optimization efforts.
What’s Wrong with this Picture?
There are a few major shortcomings of this approach that, frankly, are simply ignored by most people involved in website analytics. I don’t blame them for ignoring these downsides, because, simply put, this is the best that could be done with the technology that has been available until recently. However, when these shortcomings are understood – and overcome – digital funnel analysis will become far more effective, leading to more successful websites:
1. There is no way to know, in advance, that you are selecting the best set of funnel steps.
Think about it: all website visitors and application users do things differently, in different orders, for different reasons. It sounds like a great idea to select a number of steps and call it the conversion funnel, but, in reality, the variety in user behavior is vast. Thanks to search engines, most website visitors don’t land on the home page; they can begin their session on any page deep in the website. Once in the site, visitors can jump around and proceed along an infinite number of different paths. The steps defined for a website’s conversion funnel may have little connection with reality for the majority of website visitors!
Furthermore, there may be particular steps in the website journey that are extremely important to the conversion process that were never included in the funnels being analyzed. Even if you do a great job optimizing the steps that you are focusing on, you could very well be missing 90% of the opportunity for improving your site’s conversion rate.
2. Focusing on pageviews only tells you part of the user journey story.
While it is true that conversion funnels sometimes include one or two specific on-page actions (e.g., clicking the Add to Cart button), the vast majority of funnels analyzed today consist mainly of website pageviews. However, hiding in between all those pageviews may be numerous on-page interactions that are being completely ignored.
For example, consider the possibility that on-page user interactions – such as zooming in on product images, opening pop-ups containing shipping details, playing product videos and making size/style/color selections – are a critical part of the conversion journey for many visitors. In other words, it could be that improvements to these elements of a website could dramatically improve the business performance of a website, but because marketers are mostly focused on a series of pageviews, they are leaving a pile of money on the table.
Which brings me to my next point…
3. It has simply been much too difficult to track many individual on-page actions.
Even if you wanted to include numerous granular user actions in your funnel analysis, conventional tracking technologies make this overwhelmingly difficult. As I touched upon in my previous post, implementing tracking codes for individual website elements is complex, time-consuming and error-prone.
Making matters worse, large websites are constantly growing and changing. Very few organizations are able to keep their event tagging in step with website changes. Tracked elements may be changed or removed, omitting them from the funnel analysis without anyone being the wiser. Meanwhile, potentially important new elements may be added to the site, which aren’t tracked at all until someone realizes it and decides to add tracking for them. There is a lot of overhead involved with keeping event tracking in sync with the ongoing changes occurring with almost every large website.
4. It is impractical to maintain funnel definitions when lots of change is occurring.
Similar to the challenge of maintaining event tracking as websites are constantly changing, the rapid rate of ongoing change wreaks havoc with funnel setups: to ensure that their predefined funnels are working properly, marketers and analysts need to frequently update their funnel definitions.
Additionally, because marketers are continuously tweaking their sites and running tests to discover how to improve them, funnel analysis can be rendered inconsistent or unreliable. Even temporary changes, such as running an A/B test for a few days, can skew the results of predefined funnels.
Furthermore, while a small business owner may decide to track and analyze a single funnel for a simple website, tracking many different funnels is needed in most larger websites. In fact, I’ve seen companies that track dozens of different funnels on a single website. Trying to maintain the relevancy and value of numerous funnels in an environment of constant change is difficult, if not impossible.
Introducing Digital Conversion Funnel 2.0
The next generation of conversion funnel analysis addresses all of the above shortcomings, providing marketers and product managers with a much more powerful way to identify and understand the conversion-improvement opportunities hiding in their websites. By avoiding the limitations of conventional funnel analysis, much deeper user behavior analysis is possible, leading to countless discoveries of where to focus experimentation and optimization efforts, in order to maximize a website’s conversion potential.
Most importantly, this new approach lets the data define and refine your funnels: instead of arbitrarily deciding in advance which user actions should be linked together to define a website’s conversion funnel – and trying to maintain funnel definitions in an environment of constant change – the Digital Conversion Funnel 2.0 approach collects all user activity data and then analyzes it to determine which user actions truly make the biggest impact on conversation rates.
To become practical, this approach requires the application of recent technological innovations in four particular areas:
- The ability to easily and comprehensively track, tag and analyze all user actions, including both pageviews and every on-page user interaction, with zero coding/tagging or maintenance effort (if Human effort were required for this, it would be essentially impossible to comprehensively track frequently changing websites)
- The ability to easily and automatically analyze this large amount of granular user behavior data to determine which are the most important user actions contributing to, or hindering, conversions (even a terrific team of data scientists would be hard pressed to analyze thousands of actions within hundreds of thousands of ad hoc funnels, across unlimited dimensions, for millions of users)
- The ability for funnels to define themselves dynamically, in response to everything happening and changing in a website
- The ability to measure the conversion impact of each and every change made to a website
Imagine software that could do this, without the need for you to come up with the funnel steps to track, without the need to set up or maintain any event tracking code/tags and without the need to define or maintain funnels!
Interested in learning more? Contact us and we’ll be happy to help you gain important competitive advantage by becoming one of the first companies to implement Digital Conversion Funnel 2.0.