I've been a part of the growth field (formerly known as user acquisition/performance marketing) for my entire career, starting in ~2007 at livingsocial.com, followed by the start of my newsletter (grow.co), ratings/review site (coursereport.com), conference series (mauvegas.com), and cro/growth agency (thesistesting.com). Since 2007, I've experienced (along with many of you) major changes in the growth landscape, including:
I've shared my personal history only to put emphasis on the following point: I cannot recall an instance where our industry has been disrupted as quickly + significantly as Apple's iOS14.5 update. And I believe that a lot of brands are struggling with significantly increased CACs (50%+) over the past few weeks.
And if you are working in the growth space, you've probably had a lot of conversations like these recently:
A lot of great agencies and consultants have taken to publishing CPM benchmarks. That's certainly a quick and important way to look at the platform's auction dynamics. I don't feel as though I have enough data points to really confidently speak to macro CPM trends so instead I've pulled CPM data by account (86 in total) and presented it in a more granular fashion. I only included brands that had active campaigns from Jan - June 2020 as well as Jan - June 2021. Any brand that paused for any of those periods was excluded in this analysis, as I really wanted to be able to show the YoY CPM impact by month. This analysis includes a little less than $225m in media spend (combining H1 2020 and H1 2021).
I struggled to visualize this data and ultimately settled on the format above, though it's obviously imperfect. As you'll see in the table above, for each brand I've calculated their H1 2020 and H1 2021 (with June as a partial month) spend, and CPMs by month for Jan > June in 2020 and 2021. Lastly, I calculated the YoY CPM increase for each month.
Above is the data when averaged across the 86 accounts. We saw YoY average CPM increases of 20-25% in January and February, and a massive spike in March through June. For example, in May the average CPM increases were 80% YoY and the average CPMs themselves were up over 72% YoY. You might note that the average CPM increases don't track exactly the average CPMs themselves (though they are pretty close). That's simply a function of volatility in my data set... the key point is the average brand has seen 80%+ CPM increases YoY in May & June.
One key point is that we actually saw this spike start in March and so I think it might be as much a function of COVID/market dynamics than it is strictly speaking the impact of iOS14.5. Either way, inventory is a lot more expensive than it was a year ago during COVID!
PS: It should be mentioned that a lot of other variables are at play here, including COVID's impact on 2020 rates, individual brand's changes in campaign strategies/tactics, changes in spend levels etc etc etc.
Doing a pre vs post iOS14.5 analysis of Facebook's performance using only Facebook data is unfortunately surprisingly challenging, for a few reasons.
#1 It used to be the case that you could pull Facebook's performance data using different attribution windows. That allowed us to do analysis like this blog post where my colleague Heather looked at the same campaigns' performance using 1-day, 7-day, 28-day etc attribution windows on both a click and view basis. Now (post 14.5), you can only see performance for a campaign/ad set using the Attribution Setting set at the ad set level.
#2 Anytime you analyze multiple ad sets with different Attribution Settings Facebook will just show 0 conversions in their interfaces. Only when the entire data set is using the same Attribution Setting will Facebook show totals. It makes it effectively impossible to use the Custom Reporting tool to look at data over time, as odds are there will some different Attribution Settings in the mix at some point in time which will zero out the totals.
#3 Many (if not most) advertisers have updated their Attribution Settings after the iOS14.5 release for a few reasons, but especially because Facebook's old default attribution setting of 28-day click + 1-day view no longer exists!
In a follow up blog post (hopefully next week), I'm going to attempt a few different ways to look at pre post analysis, including using Google Analytics data as a universal constant for a few brands.
With the release of TikTok’s Creative Center, we’ve been taking a deep dive into the Top Ads section to find some trends in what is working on the platform. It’s a great resource for brainstorming new ad ideas for your ads.
At the suggestion of our Snapchat rep, we tested Snapchat's Lifestyle Categories that an account indexed strongly with to see how they performed relative to our standard broad targeting.