Confirmation Bias Makes You a Worse Advertiser

Confirmation bias makes you a worse advertiser. And the worst part is you don't even know it's happening.

Let me explain.

Confirmation bias is when you already believe something to be true, and then you only pay attention to the evidence that supports it. You ignore everything that contradicts it.

I see this all the time in Meta advertising. Here's what it looks like.

An advertiser decides that remarketing is the key to their success. They make it central to their strategy. They believe in it with every ounce of their being because of the results it produces. Or the results it APPEARS to produce. Amazing Cost Per Conversion. High Return On Ad Spend. Almost too good to be true.

They run an Advantage+ Sales Campaign and they compare the numbers. Higher Cost Per Conversion. Much lower Return on Ad Spend.

So they say, "See? Remarketing works. Advantage+ doesn’t.”

But they have no idea how much of their budget went to remarketing when they used Advantage+. They never set up Audience Segments and used the breakdown. They have no idea how many of their remarketing results come from view-through conversions. Or conversions that aren’t incremental. Because they didn’t use the Breakdown by Attribution or Compare Attribution Settings features.

Maybe they’re ignorant to these things. More often than not, they subconsciously don’t want to see the truth. That's confirmation bias. They had already made up their mind.

This is a huge problem right now because of how much Meta advertising has changed. There are strong opinions about Andromeda, about creative diversification, about Advantage+ features, about whether the algorithm can be trusted. And advertisers tend to fall into camps.

Once you've decided what you believe, you will find evidence to support it everywhere. Bad results after turning on an Advantage+ feature? "See, I knew it. I'm never “using that again.”

But you had one experience. The result may have changed if you let it run longer. Or maybe those results were based in randomness or caused by another factor. You just confirmed what you already believed.

And the thing about confirmation bias is that it feels rational. You're not making stuff up because you have real data. You have real experiences. But those experiences lack predictable and repeatable meaning. You're only telling yourself part of the story.

The advertisers who fall into this trap the hardest are the experienced ones. They've been doing this for years. They've built strategies that used to work. And when those strategies stop working, they look for reasons that don't require them to change.

I know this because it was me. I was the remarketing guy for about a decade. It was one of my favorite things to do, and I was known for creative remarketing strategies. Almost every public speaking appearance I made was on that topic because it energized me and people at it up.

So I REALLY struggled when it was time to doubt those strategies. I was questioning my entire advertising identity and worth. So I get it. No one wants to hear that the thing they've been doing for five years is no longer effective.

So instead of questioning their approach, they question the platform.

"Meta's algorithm is broken."
"Andromeda ruined my results."
"This new feature doesn't work."
“Constant outages are causing my problems.”

And they find other advertisers who feel the same way, which only reinforces the belief.

Meanwhile, the advertisers who are willing to challenge their own assumptions are the ones who adapt. They test things they don't believe in. They look at results they don't want to see. And when the data contradicts their expectations, they’re open to it. They might even believe it.

So how do you fight confirmation bias?

First, be aware that you have it. Everyone does. I do.

Second, run tests to disprove your belief, not confirm it. If you swear by your remarketing results, prove the belief is valid. Set up audience segments and run campaigns that lean entirely into algorithmic targeting. Use breakdowns by audience segments to see how much of your budget goes to remarketing already.

And scrutinize the results of your campaigns that were restricted to remarketing. How many of your conversions were view-through? How many were incremental?

Third, stop relying on single experiences to form permanent conclusions. One bad result with a feature doesn't mean the feature is bad. One good result with a strategy doesn't mean the strategy is the reason. There are so many factors impacting our performance, including randomness that we can’t ignore.

And finally, be skeptical of your own emotional reactions. When you see a result and your gut says "I knew it," stop yourself. That’s confirmation bias.

So here's the bottom of the glass.

The biggest obstacle to better Meta advertising results isn't the algorithm. It isn't a setting. It isn't even your creative, though that’s pretty important. It might be your own unwillingness to question what you believe.

The advertisers who get the best results are the ones who stay curious and stay open to being wrong. Challenge your assumptions. Test what you don't believe in. And when the data surprises you, let it.