Almost every Meta ad you launch will produce below average results

The Counterintuitive Truth About Meta Ad Performance

One of the most fascinating aspects of Meta advertising at scale is a concept that might seem counterintuitive at first:

The majority of ads you run will produce below-average results.

This insight, which I first learned from Taylor Holiday at Common Thread Collective, has been corroborated across every account we've managed.

The Power Law Distribution of Ad Performance

When you look at spend and results per ad across your Meta campaigns, you'll generally see a power law distribution. Here's what that means in practice:

1. The majority of your ads will produce below-average results.
2. A very small subset of your ads will produce such outsized results that they carry the entire account.

These outliers perform so exceptionally well in Meta auctions that they can sustain massive amounts of spend and produce results that the majority of your launched ads simply can't match.

The "Room Full of Dudes" Analogy

To illustrate this concept, let's use a slightly unconventional analogy:

Imagine you have 10 men in a room, and you're ranking them on a hotness scale from 1 to 10. Now, let's say nine of these men are 1/10, but one of them is a perfect 10/10. The average "hotness" in the room would be 1.9/10.

This means that 9 out of 10 men in the room have a below-average hotness level, while one man has a hotness level so high above the average that it pulls the overall average up.

The same effect happens in Meta ad auctions. A small subset of ads will produce the majority of the spend and results over time.

It also highlights the problem with only using the mean in your calculation of averages. When looking at averages, try to also look at the median and mode to get a more comprehensive picture.

If the calculations are vastly different, that's telling you something interesting.

Challenging Common Misconceptions

This reality runs contrary to how many people in the industry think about creative production and Meta ads. Common misconceptions include:

1. Giving every ad a "fair chance" with equal spend.
2. Making decisions based on an ad's results after a set amount of time.
3. Believing you can iterate your way to finding a winning ad by tweaking elements like background color or headline.

There's no empirical evidence I've seen that supports these approaches. It's much more likely that you'll launch 100 or 200 ads, and only one or two will produce the majority of the results.

The Impossibility of Predicting Winners

Here's the kicker: there's no way to determine the essential properties that make those one or two ads produce outlier results. It's magic, it's fairy dust, it's the crazy dynamics of Meta ad auctions.

Thinking you can somehow iterate your way towards these outlier effects misunderstands how Meta ad auctions work. This realization is crucial when developing your creative strategy, production process, and measurement approach.

Rethinking Your Ad Strategy

Given this reality, your focus should shift. Instead of trying to find the "perfect" ad through iteration, concentrate on producing as many high-quality ads as possible. This approach increases your chances of triggering those outlier events more often.

However, it's crucial to maintain a quality bar.

Increasing ad volume doesn't mean decreasing quality. Think of it like Tesla's approach to manufacturing: they've scaled up production without compromising on quality. Your goal should be to solve for volume through better "manufacturing" capabilities while maintaining consistent quality. (a comparison I again took from Taylor Holiday).

The Problem With Creative Agencies Billing You Based on a Set # Of Deliverables

This brings me to a something I see a lot of these days: creative-only agencies that charge based on a fixed number of monthly ad deliverables. If you're going to pay $5,000 a month for 10-20 ads is, you better know your margin you're going to make on average per ad.

A better approach for operators:
1. Pay for that initial set of high-quality assets once.
2. Hire a low-cost video editor to "Frankenstein" those assets into hundreds of variations.
3. Turn that initial 20-ad investment into 200 ads without the ongoing retainer.

Balancing Median Outcomes and Outliers

While chasing those outlier ads, don't forget about your median outcomes. As a good media buyer, your median ad performance should still align with your unit economics targets. You can't rely solely on those rare unicorn ads to make your overall Meta investment profitable.

Aim to set up your campaigns so that even your below-average (mean)results work with your unit economic breakdowns. Then, when you do achieve those outlier effects, they're adding pure profit to your account.

The Bottom Line

Creative production at scale, without sacrificing quality, is crucial in today's Meta advertising landscape. If you're an e-commerce owner, focus on pushing volume through your Meta ad account while understanding auction dynamics.

Most importantly, let go of the illusion that you can iterate your way to linearly improving results with each new ad. There's no empirical evidence to support this idea.

Embrace the power law distribution of ad performance, create a robust creative production process, and may the Meta auction odds be ever in your favor!

***this post was written with the help of Claude AI. I don't have the time to sit down and produce vast amounts of content. So, I record voice notes with my thoughts, transcribe them at rev.com, and then upload the transcripts to Claude AI alongside a comprehensive prompt that mimics my writing style. No, it's not perfect, but this allows you, the reader, to get the gist of what I'm saying. If you want to know more about a topic, let's speak directly.