Skip to main content

YOU HAVE TO TRAIN THE PIXEL true or false?

Are you running campaigns on Facebook? If YES, you’ll want to read this.

How many times did you hear someone say “YOU HAVE TO TRAIN THE PIXEL?”

Ever wonder what the hell that means, and/if that is the magic bullet to success on Facebook ads?

I have, because in my experience the longer an ad runs, the higher the CPA goes – which totally goes against the ‘word on the street’ (or is it a myth?) that if you train your pixel,
facebook will know who converts, optimize your targeting and serve your ad to people who will more likely convert.

There was a great discussion on this in Tim Burd’s FB Group – Facebook Ad Buyers recently, and i realized I am not alone – a ton of other people experience the same thing.

You setup 10 adsets, each different angle, you pick the very best 2-3, stop the rest and then let these top 2-3 run without touching them. CPA goes well, for a couple days, maybe
weeks, then starts going down.

Experts will say it’s ad burn out, which isn’t the case when your audience is two million, and you only showed it to 10% and frequency is 1.01 on the ad. So what is happening?

Why is CPA going up, when the angle is so awesome. Shouldn’t the pixel be working after a few hundred conversions? I mean, the word on the street says you train that pixel, it’s
going to drive you such golden traffic your CPA will actually go DOWN. NO, it doesn’t work that way it seems after running countless tests to try out this hypothesis.

SO WHAT IS REALLY HAPPENING. WHY?

According to many commenters in the FB Ad Buyers group, lets say you want people to buy something (meaning whip out and actually pay) and you have an audience of 2 million.

You select optimize for conversions > purchase. FB is going to then check the audience you targeted, separate out within that the people who bought stuff, and start showing your ad to
these people. Let’s say the 2 million audience, has 200k people in it that bought stuff. After fb algo shows your ad to this 200k, it will start showing the ad to the rest of the
audience (the other 1.8 million) and these people aren’t confirmed buyers, they might not do shit, they might leave comments/share/or like a post, or they might click but not buy. So of
course, these kind of people will NOT convert at the super awesome CPA you started seeing 48 hours into the ad campaign.

They are actually going to be harder to convince to buy something; so that is why the CPA will rise after you exhaust this buyers audience within your target audience.

What do you think?

  • Wilco Den Brok

    I think it optimizes for that audience based on two things: similarity in profile & similar in actions. That is what I think, not sure.

  • Klaus T

    I have a question.

    If I want to promote my e-commerce products, and I don’t have many purchases in the Facebook pixel, should I optimize the conversion ad for another objective (view content or ATC) instead of Purchase?

  • ktyc

    Makes sense.. Seems like constant optimizing, testing and scaling correctly is the way to go..