How do you know how well your unified search and social advertising efforts are working? To answer this question, an incremental lift analysis is a must.
What’s an Incremental Lift Test?
Incremental revenue is earnings you wouldn’t have gained without a specific campaign. An incremental lift analysis, then, assesses the average revenue from two groups:
- People exposed to certain variables within a test group (two segments)
- People in a control group
The Framework
A popular setup for an incremental analysis is testing the average revenue of the two segments of our test group. Here is an example of how you could implement an incremental lift test with Product Listing Ads on Google and Dynamic Ads on Facebook. The two formats are similar in that they are both intended for to drive sales with a product feed used to automatically generate creative.
- Control Group - Those who converted from a paid search campaign over a certain period and were not exposed to an a cross-sell ad on Facebook.
- Test Group - Those who converted from a paid search campaign and were added to a “cross-sell” audience exposed to a dynamic ad on Facebook (encouraging the purchase of a complementary item)
The hypothesis that we're testing here is that by showing a cross-sell ad of a complementary product offering on Facebook to people that converted from a paid search ad will see an increase in customer lifetime value and top line revenue. For example purposes, we could create a product set on Facebook of socks to cross-sell people who purchased shoes. We would need to create a website custom audience with a UTM parameter specifying the category of "shoes" and channel "paid search".
With the Test Group, we want to identify the average revenue impact of the group that was added to the cross-sell segment on Facebook. Then, we want to compare this impact to our control group, i.e., those who weren’t exposed to a “cross-sell” dynamic ad on Facebook after making a purchase.
The campaign should run for at least a month and you should take into consideration the sample size of your existing paid search traffic to determine which cross-sell product sets to create for a healthy population size to market to. We see a significant increase when cross-sell recent converters so make sure there is a time threshold set as to when folks see your ads.
Once the campaign has ran for a month you will want to measure average revenue per customer between the test and control group to determine what the lift is between the two groups. This should be your primary Key Performance Indicator KPI that you use to determine how to allocate your budget for additional incremental impact across Paid Search and Social. Another metric you will want to look at is Customer Lifetime Value. This metric projects how much revenue the average customer nets throughout the time period that they’re a customer. The better your customer lifetime value the lower you have to spend on acquisition costs. The calculation is simple—customer value X average lifespan. However, determining these values will greatly vary based on your type of business.
Additional Best Practices For Measuring Incremental Lift
There are several things e-commerce advertisers should consider when implementing an increment lift test. In this example, we’re using Google Analytics for attribution beyond last-click.
- When deciding how long the campaign should be, consider the average time to purchase for first-time buyers and repeat purchases. The campaign should run, at minimum, for the full purchase cycle to best determine impact.
- Simplify creative and messaging iterations, and make sure they’re consistent across both paid search and social channels.
- Deploy both the Facebook Pixel and the Google Pixel using a tag management solution. View conversion data as directional within each publisher, and compare to conversion events/goals within your analytics tool early into the campaign flight. This ensures there are no noticeable discrepancies.
- Set up revenue goals (completed purchase or pre-order request) for the products being amplified in the ads. Be sure to use a monetary value here with the funnel turned on—if you’d like to track landing page success metrics such as where people bounce within the checkout page, input URLs for each screen page the user will see.
- Ensure the UTM parameters within ads all have a standardized naming convention, so that you can run funnel reports to analyze the consumers path to conversion.
- A true measure of sales lift uses an attribution model that reveals the individual and collective contribution of each paid channel on the online (or, offline) conversion event.
- Select an attribution model within the analytics tool that your digital marketing team fully understands prior to launching campaigns. For the purpose of this analysis, the goal should be to view performance beyond post-click or post-impression metrics with a linear attribution model that gives each touchpoint along the path to conversion an equal weight.
