Measuring the Value of Non-Revenue-Based Conversion Metrics
When conversion is measured as a purchase or point of sale, it provides immediate gratification because we have a record of a sale that we can measure, with real revenue represented by the conversion event. The simple traceability of a purchase as a conversion is why we see it in so many optimization presentation decks. This is because return on investment (ROI) can be easily exposed and reliably measured within reporting, which is the essential ingredient for evangelizing and gaining momentum in your program.
However, some industries use their digital channels for purposes other than purchasing, and their conversion goals or the additional success metrics they are measuring in analytics or their optimization efforts do not link directly to a purchase or sale. They can be deriving revenue through engagement or registrations, where calculating derived revenue is not as tangible as a purchase. The most obvious example of this is in the media and entertainment sector, where accessing content (viewing an article, a video, etc.) or page consumption (time on site) are the metrics of choice. Although these content and page consumption metrics can be translated into a revenue-based number such as ad revenue, the process of calculating them differs widely across companies. Non-revenue-related conversion can be an essential metric for other industries as well. Companies in the financial services industry often evaluate their tests based on the ascribed value of new account sign-ups, whereas travel and hospitality and retail companies often have rewards program registrations as an additional conversion event or goal. High tech, business-to-business companies focus on increasing the submission of lead forms to generate more qualified leads for their sales teams.
Before diving into how our customers derive revenue value from non-purchase metrics, it’s worth noting that revenue should not be your only focus in terms of demonstrating value and deriving ROI from testing and optimization. Testing to learn also provides valuable market insights by putting assumptions or well-worn beliefs to the test, such as the persistence or effectiveness of a content carousel versus targeting specific offers, based on data captured from customer preferences relative to conversion. Also, testing to learn can quantify the effect of fresh ideas around customer experience and engagement by using testing and optimization as an exploratory mechanism. This approach allows us to deconstruct and establish the reasoning behind our existing targeting and personalization decisions.
A/B and multivariate testing, in addition to utilizing an algorithm for automated personalization, are also helpful in reducing time and resources invested in creative and development resource allocation, as the approaches for personalization across locations are more clearly defined. This eliminates the endless variations of creative content previously needed to assess success and can now be generated and maintained by a non-IT resource, which means less reliance on development resources. A more relevant customer experience also reduces call and chat center costs, which results in cost savings for the company. To truly understand the impact of a test optimization program, both testing to learn and time/cost savings should be evaluated and evangelized as derived value in addition to the conversion or revenue lift produced.
OK, let’s jump back to evaluating non-conversion-based success metrics. When looking at visitor engagement, time on site (page consumption), or registration/subscription as a metric, it’s important to first determine some basics. What pages are considered high value in terms of generating the most revenue? What pages have the highest traffic? How often do visitors return? What pages are notorious for visitor abandonment? Once a baseline understanding of the dynamics of your content are understood, you can use this information to your benefit. You can test experiences by pushing visitors to these high value pages based on different paths through your content. Test several contenders in high traffic locations and push visitors from high traffic to high value pages. Identify goals for distinguishing between infrequent and loyal customers and focus on retargeting personalized content for return visitors. Finally, design tests to re-engage abandoners at point of abandonment.
Alongside a solid test strategy, it’s important to quantify the value or revenue based on the relative value and volume of pages consumed. An easy first step is to apply an approximate value, such as equal average revenue value, to all pages and begin to get a rough sense for overall revenue per consumed pages. This rough approximation demonstrates approximate values for overall revenue and reveals potential high value opportunities based on targeting potential (i.e., a location where a large visitor segment is dropping off, but could be re-engaged and pushed to an additional page). Another approach is to apply the revenue value or approximate potential revenue derived from an ad placement on a specific page as a gauge for revenue derived through test outcomes. If this number varies daily, weekly, or monthly, many customers will apply a scoring matrix to their tests. This is where pages are scored based on relative value to each other, with their score being applied to current page revenue based on a given window in the test process. This is also valuable in estimating lift derived through testing and improving content or pages consumed, leading up to a lead form submission, account signup, or rewards program registration.
One of the benefits of this method is that custom scoring can be set up as a success metric within a test or targeted activity in the Target solution. This makes measuring and generating reports on these metrics much easier and allows for ROI to be assessed on more than just a purchase. Keep in mind this scoring technique can be applied to more than just page consumption; we can apply the same system to products, registrations, or lead forms based on the value of signing up or registering for a specific product, service, or offer.
We’re continuing to develop ways to expose and measure ROI and the revenue generated within the solution using new features. You are now able to ascribe a specific revenue value to your conversion event so revenue generated during the test can be viewed in reports and seen in aggregate across your activities. And we’ll be adding more soon, all while ensuring that the metrics we expose have the same statistical rigor as the confidence levels in our conversion lift/results.
So, let’s continue to measure and evangelize all of the value our companies can realize through testing and targeting different variations in content and experiences. This data can be used to further tailor your test program, even in the cases where revenue and value need to be derived through a matrix or formula.