The Secret to Finding Winning Ads

One of the brands I make ads for recently went from:

• $6,888/week
• 1 month later: $43,417/week
• 2 months later: $64,355/week

That's over $250k a month in sales driven primarily by Meta ads.

But growth has recently slowed.

This is due to a mixture of factors, but the only one I have some control over is the ads.

So I've been back to the drawing board thinking of the best way to make the ads that are going to help scale this to $500k a month and beyond.

To do this, I've gone back to first principles. I've crystallised my entire system for finding winning ads and thought I'd share the most important part with you today.

When most people run Meta ads they:

• Waste money on random tests that don't have a clear goal aside from "let's test this idea I had last night".
• Have no idea whether their test has a high or low probability of success, and invest the same amount of money and time into both.
• Have no creative strategy system for finding new winning angles and capitalising on the learnings they've made so far.

So to solve ALL of this, I want you to realise that finding winning ads is less about 'just making better ads' and more about finding winning angles, then making lots of ads to reap the rewards of those angles.

So what we really have is a 'search problem'.

Anyone can make lots of ads. Not everyone can make the right ads with the right angle. The real alpha is being able to do this with AS LITTLE wasted resource as possible. The more efficient you are, the less money you waste on losing angles, and the more you spend on winning ones.

To create a system for this, I've borrowed a few frameworks from the mathematician Bernard Koopman. During WWII, he developed breakthrough search frameworks for the U.S. Navy to help them find submarines and downed aircraft. He's regarded as the father of modern search theory.

Imagine finding winning ads is like mining for gold in a forest. To find the pockets of gold as fast as possible, there are three types of search:

1. Parallel Search

If you have NO knowledge where the gold is, your best bet is a broad search of an area in the pattern of parallel lawn mowers. The goal here is not a 'deep search'. Instead you want each 'probe' to search as wide a radius as possible.

2. Spiral Search

If you have info that the treasure is near a specific point, your best bet is to have your probes spiralling outwards from that point.

3. Optimum Effort Search

Map out the entire area that you can search. Assign a probability to each section. Then focus all energy on the area of highest probability. If the search fails, update the probabilities of every area and move to the next highest probability spot.

We can apply these SAME strategies to find winning ads. I've adjusted it slightly to the context of Meta ads, but the principles are the same. Essentially you want to categorise every ad as either a miner, a spiral or a probe.

1. Probes

Ads where you're testing a completely new angle or avatar. If your primary audience for an acne product has been teenagers, you test running ads to pregnant women with acne breakouts.

2. Spirals

Where you've had some positive data from your probe, but don't yet have a strong winner. You're testing around an angle that you think has promise but haven't hit the spot yet.

3. Mining

You've got a winning angle. Now your job is to extract as much value from that angle as is humanly possible. Different messaging, formats, hooks etc, all hitting the same angle.

But we don't stop here. Next we can steal Google's 70-20-10 innovation rule to really make this efficient. Eric Schmidt, the former CEO of Google, often credits the company's success to this rule. Essentially it says:

• 70% of efforts should be devoted to core business activities.
• 20% should be focused on projects adjacent or related to the core business.
• 10% should be dedicated to completely unrelated or experimental projects.

So when it comes to making winning ads:

• 70% of your ads should be on Miners.
• 20% of your ads should be on Spirals.
• 10% of your ads should be Probes.

If you don't have any winning angles yet, then use 80% of your energy on Spirals and 20% on Probes.

Remarkably simple, but I find this is an extremely useful framework for instantly gauging whether your ads are being made in the right area.

Some businesses make the mistake of only spiralling and probing, always looking for new angles and never taking advantage of their current winning ones. Others make the mistake of stopping their search after finding one winning angle. Unless the market for that angle is huge, they'll eventually hit a plateau and wonder why they are no longer scaling.

This is why the best ad strategy fits into the 70-20-10 rule.

Happy scaling,

Tatsuki Thomas