If you’re here, you already know what the Voter Formation is about: we’re a 501c(3) here to modernize how online campaigns register and mobilize under-represented voters (especially people of color). So this year, we ran a big experiment to see how we could better reach those voters, starting with Latinas in Texas and Black men in Virginia. And, we’ll be honest: our findings surprised us.
We knew from the start that our 2021 campaigns weren’t likely to be as cost-efficient as previous, on-cycle campaigns in terms of immediate costs per conversion. A key component of our program was centered around driving users to fill out a form and “make a plan to vote,” reinforcing that commitment to actually vote by Election Day. We expected lower engagement, and we knew that the release of Apple’s new privacy framework, called App Tracking Transparency (ATT), had the potential to fundamentally change the media landscape.
And while we still don’t have enough information to know exactly how voting behavior was impacted, we do know that engagement was lower, and that ATT changed everything.
App Tracking Transparency 101
Let’s back up: in April, Apple released iOS 14.5, a software update that included a big new privacy measure. With the update, Apple required apps to ask users for permission before tracking their activity. We can’t emphasize enough how much this disrupted the advertising models of platforms like Facebook and Snapchat that rely on apps to reach most of their users - without access to user data across multiple platforms, it’s a lot more difficult for any one advertiser to be sure that they’re targeting the right people.
But that's not all: not only was it more difficult to target users generally, it also appears that Facebook was intentionally suppressing the delivery of ads to iOS devices.
We can start by comparing our 2020 Facebook Ads Conversion campaigns to those we ran in 2021: last year, before Apple released iOS 14.5, Facebook Ads delivered 63.6% of impressions to iOS devices. This year, Facebook Ads delivered only 36.8% of impressions to iOS devices. That’s a 42% decrease in conversion campaign impressions to iOS devices.
Diving into ad delivery by device type and campaign objective, the extent to which Facebook suppressed conversion campaign delivery to iOS devices becomes crystal clear: of all impressions delivered to our iOS audience, Facebook delivered only 5.9% via a conversion campaign, compared to 20.2% for our non-iOS audience — a nearly 71% difference in conversion campaign impressions by device type.
Because Facebook delivered a disproportionate number of Conversion campaign impressions to our non-iOS audience, demand for ad space was
higher on these devices, resulting in conversion campaign delivery costs that were 44% higher on non-iOS compared to iOS devices.
These results strongly suggest that iOS 14.5 is the cause of the disproportionate difference in delivery by device type among conversion campaigns. Facebook has suppressed delivery to devices that are more likely to block its ability to track the conversion action.
What does that mean going forward?
iOS 14.5 has made it more difficult and expensive for us to drive conversions, point blank. And Apple’s update is likely only the first of many privacy changes coming to the media landscape. It seems safe to assume that the average cost per conversion will remain higher than it was before iOS 14.5.
Given the racism and biases that are built into the data we rely on to run these programs, it’s difficult for us to reach a significant portion of our audience to begin with. If we were to build a Facebook strategy centered around driving conversions, no matter the cost, we’d just reduce our reach on a platform with a wide user base within our target audience, which isn’t an acceptable alternative.
Abandoning conversions altogether isn’t the right approach, either, but we can no longer rely on Facebook Conversion campaigns to drive them at scale. We have a lot of questions we need to answer before we find the best way to approach the new and changing media landscape. Here’s where we’re starting:
-
Could we scale the incorporation of more native conversion points, like Facebook’s Lead Gen forms, across other platforms? How would this impact our ability to measure our program? Would the potential benefits outweigh the downsides of driving less traffic to our own website?
-
How can we more intentionally “nurture” our audiences to conversion, using the corporate marketing funnel as a blueprint? What messages would work for different audiences, and how and when would we test that?
-
How might some of the tools platforms are building in response to iOS 14.5, like Facebook’s Conversion API, impact our ability to drive more efficient conversions? Are the time and resources required to set this up worth it?
-
What other platforms should we test, and what kind of engagement or conversion rates would they generate among different audiences?