Why Your Performance Goal Changes Everything

In a prior episode, I talked about how the algorithm works, and this is critical to understand. So go back and listen if you haven't yet.

The short version is this: the algorithm is literal. You define what you want, and the algorithm tries to get you that action.

Today, we’re going to dig into how to actually go about this.

The Role of the Performance Goal

Your performance goal, which is set in the ad set, defines what you want. It controls both ad delivery and optimization.

Examples of performance goals include:

• Maximize number of conversions

• Maximize value of conversions

• Maximize link clicks

• Maximize landing page views

• Maximize through play views

There were 21 different performance goals the last time I counted, and there could be more now.

Since this setting defines what you want and how your ads are delivered, its importance cannot be overstated.

Why This Matters in a World of Algorithmic Targeting

Targeting is mostly algorithmic now. If Advantage+ Audience is turned on, nearly everything you input is treated as a suggestion.

Even if you turn Advantage+ Audience off, detailed targeting and lookalike audiences are often expanded, meaning they are treated as suggestions too.

So how does algorithmic targeting work? Meta shows your ads to people who are most likely to perform the action you specify.

And how is that action defined? By your performance goal.

Your performance goal effectively acts as your targeting. It may not fully define it, but it plays a major role.

You might feel like you’ve lost all control over targeting, but that’s not entirely true. You will reach different people depending on your performance goal, even if you can’t see the details of how that targeting works.

That’s why your performance goal can lead to either high-quality or low-quality results, depending on your selection.

Example: Maximize Link Clicks

Let’s say your performance goal is maximize link clicks.

The algorithm says, “Great, let’s get you as many link clicks as possible.”

It might start by showing your ad to a few people who have clicked before or visited your website. But its intent is not to find buyers. It doesn’t need to show your ad to people likely to become customers.

Reaching those users could be more expensive than reaching someone else.

So where does it go? The algorithm might lean heavily into the Audience Network placement. It doesn’t matter what your business is or what’s in your ad. Audience Network is full of cheap clicks.

Sure, many are accidental. Some may even be click fraud before detection. But that doesn’t matter to the algorithm. You asked for link clicks, and that’s what it’s going to deliver.

Even if you remove Audience Network, the problem doesn’t disappear.

The algorithm is still focused solely on finding cheap clicks, and it will find them. Some people click on everything. Some are confused. Some may not even be people at all—just bots or spam.

And this kind of thing only seems to be a problem when optimizing for low-intent actions. You don’t get “cheap purchases.”

Example: Maximize Conversions (Purchase)

Now let’s change the performance goal to maximize conversions, and choose purchase as the conversion event.

At that point, everything changes.

The algorithm still wants to get you as many of the desired action as possible. But the weaknesses it previously exploited are no longer available.

There is no such thing as a cheap or accidental purchase from a placement.

Your ad probably won’t even appear on Audience Network. It’s strange how that works.

Without touching your targeting settings, your audience changes. A larger share of your budget will now go toward remarketing—prior customers, your email list, website visitors, and people who have engaged with your ads.

Why? Because Meta knows those users are the most likely to buy.

You’ve changed your targeting simply by changing your performance goal.

Change the Conversion Event: Leads

Now, let’s stick with maximize conversions, but change the conversion event to leads.

Your targeting changes again.

You’ll still get more remarketing than with a link click campaign, but less than with a purchase campaign.

Why? Because leads can be gamed. The algorithm can find people who will submit forms, even if they’re not quality leads.

You want cheap, high-quality leads, but the algorithm doesn't necessarily know that.

For some reason, I’ve seen the algorithm favor people over age 65 when optimizing for leads. Unless I set an age maximum, a huge percentage of my budget goes to that age group.

Final Thoughts

Your performance goal is incredibly important.

Do not try to game the system by optimizing for link clicks, video views, or other low-quality actions with the hope of getting cheap purchases.

That rarely works.

The same goes for optimizing for add to cart, initiate checkout, or view content.

You may have lost some control over explicit targeting, but your performance goal remains one of the most powerful levers you have.

So do not minimize it.

And do not complain about low-quality results when using a performance goal that doesn’t take quality into account.

I wish Meta would provide a way to optimize for high-quality top-of-funnel actions, but that’s a separate discussion.

When possible, use maximize conversions with the purchase event.

If you use another goal, understand the risks and do what you can to mitigate them.

I wrote a blog post about inflated results and how to address them. It includes a full roadmap based on the various performance goals.

You can read it at jonloomer.com/inflated.