Case Studies
Facebook Ads by Wix | Reduce Churn
Facebook is one of the best advertising platforms in the world. It's also one of the most complicated. To help our users with this important tool, Wix developers created a machine-learning algorithm that does the work for them.
I worked on this product for 2 years and was involved in every aspect of it, from the very first wireframe to the most recent feature release. This is just one example of the many product optimizations and new features we did during those 2 years.
Challenge
One of our main KPIs is reducing the churn rate of "successful" campaigns, defined as campaigns with 300%+ return on ad spend (ROAS). Another main KPI is to increase the percentage of running campaigns that are successful (again, 300%+ ROAS).
After talking with users, we learned that there were 2 clear groups who needed more info:
1. Users who were ending successful campaigns without even realizing that they were successful.
2. Users who were ending unsuccessful campaigns too early, so the AI algorithm didn't have enough time to learn.
Solution
We created 3 different messages, each tailored to how long the user had been running their campaign and how successful it had been thus far, measured by return on ad spend.
-
Less than 10 weeks and under 300% ROAS
-
300% or higher ROAS
-
More than 10 weeks and under 300% ROAS
We also added a feedback form, so we could better understand why users were leaving.
Results
There was a 12% decrease in the number of users canceling their campaigns from this flow.
Those who did end their campaign left us incredibly valuable feedback through the feedback form. It helped us identify a technical bug we weren't aware of, it led us to do an A/B test on a modal we hadn't thought about in months, and it helped us prioritize our next feature releases.
My Contribution
I worked very closely with the product manager (PM) and UX designer on these modals:
-
Helped decide the layout and design of these modals, trying out several different things with the designer
-
Involved in the discussion about which icons to use for the modals, as the icons very directly affected the tone of my text
-
Contributed to conversations regarding the order of the CTAs, the color of the CTAs, and overall behavior of the CTAs
-
Determined which options we would present to the user on the feedback form
As always, I was also responsible for QA'ing the content, sending it to localization, and overseeing the resolution of issues (i.e., responsive buttons and size of modals) raised during the localization process.
