After iOS 14, businesses stopped trusting the numbers they were seeing in Facebook. Understandably, brands started looking for a third-party tool to give them the clarity they once relied on Facebook to provide.
Enter: Multi-touch attribution tools.
What started out as a helpful way to understand customer metrics has turned into over-reliance, and dare we say, an over-hyped understanding of the data presented.
Which raises the question: Is there such a thing as too much data?
Despite CTC’s data-centric reputation, our answer is a resounding “Absolutely.”
As we continue to work with third-party attribution tools like Rockerbox and Triple Whale, we’re seeing some discrepancies and pitfalls to be wary of.
Yes, the reporting and visibility on the data has changed. But that’s no reason to put all your chips down on tools that promise to give that visibility back. That’s because there’s a difference between measurement and decision-making.
The main purpose of attribution tools is to assign credit for a sale to a specific channel within the marketing or media mix (i.e. Facebook, Google, TikTok, etc.).
These tools serve two main purposes for DTC businesses:
Reporting & analytics within these platforms are meant to show you the full customer journey, rather than just consumer behavior as it pertains to a single channel.
Attribution requires some sort of modeled framework (each tool has different ones) to understand how much each channel contributes to parts of that customer journey.
Practically, media buyers analyze the data reported in these platforms to provide strategies for enhanced customer acquisition.
Multi-touch attribution applies different weights to each channel in the marketing mix. Ultimately, all channels add up to 100% of the businesses actual revenue outcome.
Typically, performance will look a lot worse than what Facebook is showing.
Anecdotally, we’ve seen platforms like TikTok and Google (branded search campaigns in particular) seemingly perform higher than expected.
So what’s going on here?
Here’s an illustration: You see an ad on Instagram, click on it, browse a little bit, but you don't convert. Later on that evening, you Google that brand, branded search comes up, and now you buy …
Both Facebook and Google will take credit for that purchase.
This double-dipping is exactly what multi-touch attribution tools try to solve — and retailers try to use these insights to better understand where to put their marketing dollars.
In the above example, the multi-touch attribution tool might report that Google's getting 75% of the credit for that purchase and Facebook's getting 25%.
So … you just cut the efficiency of Facebook by three quarters. Is that really accurate?
When it comes to multi-touch attribution, you don't get any of those subsequent touches — and you don't get the conversion — without the first touch.
And the data backs that up: across the board, a decrease in Facebook spend correlates with a decrease in overall revenue.
“Is this exact number in Facebook accurate?” is the wrong question. Because even if Google is “75% responsible” for a sale, that sale would never have happened without that initial 25% from Facebook.
Beware of getting so caught up in the discrepancy between the two numbers that you pull back on Facebook.
There’s another major flaw with multi-touch attribution tools — none of them pass their information back to Facebook. And that means that Facebook can’t price your ad or optimize against that data.
Inevitably, you'll start to see your CPMs go up, since you're telling Facebook to optimize for people that its own data shows to not be profitable. This perpetuates the problem, creating even wider discrepancies in your reporting.
So, what should you do instead?
Instead of blindly trusting the attribution tool, compare the attribution reports to channel and store data.
Here’s what we’ve found running comparisons of our own …
There’s, at best, a loose correlation between attribution tools and channel metrics.
When we talk about correlation, we’re simply looking at how these metrics move with one another. If attribution reporting goes up, Facebook reporting should too — even if it's by different margins.
In our analysis, we’ve mostly found a loose correlation between third-party reporting and channel metrics. This tells us these data points are not tied closely enough to make decisions in Facebook.
For instance: Some days, a 3.0 ROAS in this attribution tool might equal a 1.5 in Facebook. But other days, it’s a 2.0, or a 3.0 , or an 0.87.
This makes it especially hard to set goals in these attribution tools — if changing things in Facebook doesn’t change the attribution data in any predictable way, your only real option is to close your eyes and hope for the best.
A third flaw in multi-touch attribution tools: they don’t take LTV into consideration.
Why is LTV so important to attribution?
We’ll use a real-life scenario from a brand that used a multi-touch attribution tool.
The tool reported that Google traffic was inexpensive, resulting in significantly better CAC and ROAS.
At the same time, the tool reported terrible CAC and ROAS on Facebook, leading the team to conclude, reasonably, that Facebook was way too freaking expensive (their technical term, of course).
So they shifted their budget away from Facebook and towards Google, hoping to see improved MER across the board.
A month later, MER had gotten worse. When the team looked at LTV by channel, they found that Google’s LTV was half that of Facebook. They ended up relocating spend back to Facebook to acquire more high value customers.
Moral of the story: Each data point is only one piece of the puzzle (or rather, one piece of CTC’s “Hierarchy of Metrics”). You have to use all of the metrics in relation to each other to provide context for each individual performance indicator.
Since multi-touch attribution tools can’t provide a full picture of the customer journey — ironic, given their promise to shed light on the full funnel — you’ll find yourself optimizing for cheap traffic without considering the long term outcome.
We recently partnered with Retina.ai, an AI platform that solves for all three of these problems:
Despite the skepticism around Facebook data, brands were quick to flock to these attribution tools as the ultimate source of truth. But we should always be weary with data.
We don’t mean to completely dismiss these tools. It’s not unreasonable to want to use or rely on them. Especially, if you don't already have a platform that allows you to see all your metrics in one place.
Digest the numbers with an interest and a curiosity. Data is informative, but ultimately, platform best practices are the way drive growth on individual channels.
We won't send spam. Unsubscribe at any time.
As a Sr. Growth Strategist at Common Thread Collective, Luke Austin leads some of the most exciting $100M+ consumer ecommerce brands in the industry. Connect with him on Twitter about DTC growth, innovating margins, or the future of NFTs and Web3.