People still don't understand the breakdown effect

Every day a new media buyer is birthed into the DTC industry without knowledge of the breakdown effect in Meta Ads

This will be a short post.

This isn't a new concept. Meta has documented it comprehensively for years.

Yet buyers & those looking into their Meta accounts still don't understand.

Let's make it very simple. Looking at the Meta Ad performance over time of two ads red and green.

Two ads run over any given amount of time.

Then, at some point in time, media buyer logs into the ad dashboard. (dotted line)

Media buyer or marketing manager looks in and says "oh wow, look! The Green ad is performing better than the Red ad"

"Let's turn off the red ad and then spend more on the green ad."

(obviously the media buyer isn't privvy to the entire line on the chart).

But what's wrong with this?

Well, you, the media buyer, are making a decision about what to do based off of past performance, past outcomes.

Meta, however, is making decisions about modeled, predicted future performance.

By using the ad dashboard and looking at past ad performance, you are making the wrong decision for future growth. The graph obviously illustrates this perfectly. The green ad that you decided to keep on loses all its performance and you've robbed the red ad from its future performance.

Some ads are able to support higher spend and perform better in future ad auctions that have yet to happen. You, the media buyer, don't know that. And when you look at "ROAS over the last 7-14 days" and use that as your decision making barometer, you end up making harmful decisions to your bottom line.

Meta's ad auction delivers ads based on expected future results.

Media buyers will look at ad or adset breakdowns and be like:

"why did meta spend 5x more on this ad (Ad A), when this other ad over here (Ad B) had a 40% higher ROAS?"

This is the breakdown effect in action. The ad that recieved more spend was able to support more scale. In the counterfactual situation, where Ad B had received the same amount of spend as Ad A, the ROAS would've been lower than the initial ROAS on Ad A.

This is a difficult concept to grasp.

If you're still confused Dara Denney gives a very good breakdown of this in a video here.

What does it mean? Simply, don't make decisions in your ad account based on what happened in the past. Instead, determine the outcome of what you want to happen (like "hey Meta, I need to acquire a customer at a target CPA of $60") and then allow Meta to make the decisions of which ad should be delivered in any given auction.

Because Meta isn't deciding this based on a "trailing 7-14 day ROAS." Their tools are a lot more sophisticated than that.

You're not going to beat that.