Simple Isn't Always Better for Meta Ads

So if you have been listening to the podcast for a while, you know all about how I recommend a simplified approach to campaign construction.

But I understand the counterarguments. There are exceptions where simple is not always better.

What matters is that you prioritize a simplified approach while also understanding when exceptions make sense.

Let me explain.

In a perfect world, you would have one campaign and one ad set.

It is a perfect world because all of your budget, focus, and resources go into that single ad set. There are no concerns about watering down your budget by splitting it between multiple campaigns and ad sets.

There is no auction overlap where you are trying to reach the same people with multiple ad sets and driving up costs. There is no audience segmentation that prevents Meta from optimizing for the best audience.

It is the most efficient approach, and it gives you the best chance of exiting the learning phase. You limit or eliminate problems related to low volume that come from overcomplicating things.

But maybe you want to build your email list and sell products, and you have the budget to sustain both.

Maybe you have multiple products that are business priorities and each requires dedicated budget.

Maybe you are part of a large company with multiple locations, and all of those locations need to get some love.

In those cases, relying on a single campaign and a single ad set is not reasonable or possible. Those are exceptions.

Even then, you should still prioritize simplicity. Avoid unnecessary complexity whenever possible. Know that each additional campaign or ad set can contribute to the problems we are trying to avoid.

Now let’s talk about targeting.

I strongly advise tossing aside old school approaches to complex targeting from years past.

If you are using detailed targeting and lookalike audiences, those inputs are almost always used as suggestions. It is questionable whether they make any impact at all, so there is no reason to obsess over them or create separate ad sets based on those inputs.

Algorithmic targeting cannot be avoided in most cases, so stop fighting it.

There is also rarely a reason for remarketing now, especially general remarketing where you target all website visitors or your email list. Meta already prioritizes these people when you use broad targeting, and you can prove it with sales campaigns and audience segments.

You also rarely need to restrict by age or gender. Even if you know your ideal customer is a woman between the ages of 35 and 54, Meta will learn that when you optimize for purchases and adjust delivery accordingly.

So targeting should generally stay simple. Avoid unnecessary complexity that often just makes results worse.

But again, there are exceptions.

Maybe Meta is dedicating too much spend to remarketing when using algorithmic targeting, and you want to control that spend. To do that, you would need separate ad sets and an ad set spending limit for remarketing.

Maybe you are getting cheap leads from people over the age of 65, and Meta is spending a large percentage of your budget there. This has happened to me, and I know very few of those people end up buying.

You could restrict targeting by age, but you could also use value rules.

Maybe you are running engagement ads to increase views of your Reels, and your service helps women starting a business. Because the ads are optimized for engagement, you get lots of empty engagement from men, and Meta starts targeting them more.

You could restrict targeting by gender, or again, use value rules to control it.

What about placements?

In most cases, you should keep things simple and use Advantage Plus placements.

But there are exceptions when you optimize for actions other than conversions.

Audience Network can lead to cheap clicks when you optimize for clicks. Rewarded video can lead to cheap views when you optimize for ThruPlay. Facebook Reels can lead to cheap impressions when you optimize for reach.

If you optimize for any of those actions, you should absolutely consider removing those placements. Otherwise, you can often use value rules to control this behavior instead.

So here is the bottom of the glass.

You should prioritize a simplified approach, but there will always be exceptions where adding complexity makes sense.

The key is threading the needle between too simple and too complicated.

Start as simple as possible. Only add complexity when it solves a known problem.