Performance Marketing Reboot - 8.28.24
A Marketing Newsletter for Performance Marketers Converting to Marketers™
As a recovering performance marketer™, sometimes I do truly enjoy dipping my toes back into the DEPTHS of “performance marketing”. That has been my life the past few weeks as we have uncovered some anomalous behavior across dozens of Meta accounts. In this month’s newsletter, I’ll give you the 10k foot view of the output from 20+ of hours of digging into Meta performance. However, I will also try to bring a more holistic marketing lens to the problem we have uncovered - i.e. are we solving a channel specific problem focused on efficiency or are we solving a marketing problem focusing on effectiveness.
Let’s dive in.
Highlights:
How Amazon and Walmart are skewing ecommerce growth (and what it means for your brand)
Age distribution optimization on Meta is a massive unlock + some immediate recommended actions for your teams
The Harris Campaign’s $370M push, and other election spending narratives
Unedited/Unpolished UGC content is IN
Post I like:
This post from the Chief Strategy Officer at Publicis (Jason Goldberg). A good example of how averages can be misleading. Here’s why the ecomm growth story is actually different from the headlines:
Amazon represents 40% of ecommerce sales - growing at 8-10%
Walmart represents 7-10% of ecommerce sales - growing at 20%+
If total ecomm growth rate is 6.6%, then the implied growth rate of all other ecomm is ~2% (I’m happy to share my algebra to get to this - shoutout Ms. Marion from 7th grade).
There are a TON of implications to this, but two of my favorite takeaways from our team:
You have to meet customers where they are - or you really need to understand the tradeoffs and have strong reasons not to. Oh, and you need a media and measurement strategy that supports this model.
Customers are not robots (!!!) - they will buy across multiple storefronts, and their behaviors (and loyalty) will vary over time. This makes both physical and mental availability extremely important (be top of mind, be shoppable where they are).
Test I’m Excited About:
This is a big deal. Meta might be sub-optimally delivering your ads based on age. Why? Well, it’s likely because older people click more, and Meta uses clicks as a signal for optimization. (side note: clicks are not as important as you might think). The impact of this is meaningful - not only are you potentially delivering to sub-optimal audiences that convert less/are less likely to be buyers of your brands, but you are likely seeing much higher CPM’s, reducing your reach over time and thus reducing your overall marketing effectiveness.
Our rabbit hole project started with one client, where over a short period we watched sales fall off dramatically. After hours of digging, we eventually found that our budget had slowly started skewing into older age groups (we target 18+, “best practice”).
Here is a sample of the data we found:
CPM’s for 65+ are 6x that of the 25-34 cohort
Click Through Rate for the 65+ bucket is 3x vs the 25-34 cohort
What makes this even more painful is how this backs out into overall efficiency. In this case, Meta started pushing > 50% of our spend into 55+ age groups, but these age groups were producing CAC’s 3-6x that of younger demos. Woof.
What did we do?
I’ve previously mentioned Meta’s Bid Multiplier in this newsletter (we have it at New Engen through limited API access), and it continues to be a remarkably valuable tool for our teams to optimize meta performance.
The below tables shows the same client’s distribution of spend and conversion volume across age groups. We setup bid multiplier rules on 8/10 after determining Meta was over-indexing into the 55+ age group. The correction was almost instantaneous - the 35-44 age group went from 18% to 40% of spend and our overall CAC cut in half. We have some fine-tuning to do, but we stopped the bleeding.
Here’s what you can do:
Action 1: Pull your Age performance, including % of spend by age bracket and % of conversions/revenue by age bracket. You can create a view like this.
Some variance is normal, and of course sub-par performance at an age level may speak to misaligned creative or site experience (assuming your product is a fit). However, look for extreme variance where you are clearly wasting 10, 15, 20% of your budget.
Action 2:
If you’re using standard campaigns (non ASC), change your age targeting and opt out of advantage audiences.
If you’re using ASC, you’ll need access to something like Bid Multiplier. Meta’s new conversion rules are coming soon, but are not rolled out to everyone yet.
There are a number of other changes coming to Meta (new attribution, specifically) that will further contribute to these types of outcomes. This topic has garnered quite a bit of attention over the last few weeks. Here is another article I recommend, this one from Steven Johnson on Measured’s Blog.
Macro Observations:
The election is the obvious macro headline at the moment, and our team at New Engen has built a slick, fairly real-time dashboard tracking digital election spending by platform.
One things really jump out at the moment: Digital spending (excluding CTV) is way down vs. 2020…so far. Our data suggests spend in mid-August was down by 50%.
Why is that and what does it mean?
A Delayed Ramp: With Candidate Harris’s late Entry and no primary, we’re seeing a much delayed acceleration in spending. But the acceleration has begun and the Harris campaign specifically has cited they will spend $370million between Labor Day and Election Day.
CTV and Linear TV the Focus: the Harris camp specifically has stated they will not be advertising on social media (our data suggest otherwise, but campaign level data is noisy). Expect the bulk of their spending to go toward CTV. The thing to watch here will be CPM’s across platforms - will social CPM’s hold steady? YouTube? We’ll have updates on this in our dashboard.
Within non-TV Digital, YouTube leads the way: YouTube, campaign to date, represents about 30-40% of digital spend.
Reading:
Why Creators Have Stopped Editing Their Content: classic trend reversal here, and these seem to be happening faster and faster. The lesson for us marketers here is 1) always be on the lookout for the counter trend but also more tactically 2) don’t worry too much about super polished content, especially when it comes to UGC/creator content.
7 Things I love about Facebook Ads: Algorithms are crazy, and because of some of my posts/interactions about ASC and Meta age groups, my entire feed (and several coworkers’s) is now Meta-specific growth content. This is a good piece that specifically highlights two cool concepts - the first is around using engaged 1-day view for attribution and the second is around awareness optimization for remarketing (known customers, this is important). They claim to have run 5 lift tests where this approach won out 80% of the time vs. conversion optimization.
Fewer Teens Want to Drive. It’s Changing How they Spend: these types of cultural shifts are actually quite meaningful for brands - this is the type of information that great strategy needs to capitalize on. The headline stat: “The percentage of 19-year-olds with a driver’s license dropped steadily from 87.3% in 1983 to 68.7% in 2022. This is WSJ paywalled - if you want a copy, let me know.
Inside GAP CEO Richard Dickson’s Plan To Get Another Legacy Brand Back on its Toes: a longer, but worthwhile read on strategy from someone with a track record of strategic breakthroughs. A few good reminders that marketing is not just advertising and how innovative thinking is both stifled and promoted.
Thank you for reading. Leave comments or send me a message if you’d like to chat.
Kevin