My free AI ad generator killed it on socials recently.

3,000+ comments, 2,000+ likes, and THOUSANDS of new followers later…

Here’s my apology:

(Comment anything and I'll send it to you - must be following so I can DM you) A few weeks ago -

I did a simple give away on here for a custom GPT trained on $500M worth of Meta ad spend learnings.

Trained to write top-performing ad scripts for ANY brand.

What I didn’t expect?

🤯 I made this full commercial using only AI.

No film crew. Just 4 pictures, 4o, Kling, and a little editing.

Here’s how I pulled it off—and how you can too 👇

Bonus: Comment 'Prompt' and I'll DM you my private CustomGPT I used to produce the source images. It all started with these 4 simple source images:

Models holding the client’s product, and a close up of the product itself.

That’s all. No storyboard. No script.

Just a vision—and a CustomGPT I built to write killer image prompts.

The prompt you use in ChatGPT is the most important thing to get right

So I built a CustomGPT that creates a perfect prompt for image generation

Specifically, realistic UGC featuring your product

Here's how it works 👇

PS - Reply 'GPT' to this post and I'll send you the link First of all, the secret sauce.

When creating a meaningful prompt, you need to be very specific. So, I created a comprehensive set of instructions for a Custom GPT to save some leg work

Using this Custom GPT will save you having to prompt every time you want to produce an image

This is insane 🤯

I built this entire ad using AI - from the script to the animation and voice over.

Here is the exact 6-step system I used to do it 👇

Bonus: read below to find out how to get my CustomGPT that writes banger ad scripts in your DMs 👀 1️⃣ ChatGPT needs to understand your brand. This is a hack I learned from @apollonator3000.

Give ChatGPT images of your brand and ask it to create a brand profile JSON. This outlines all important contextual info it needs.

Paste this into a text file and save it for later.

In 2022 I managed nearly $100MM in ad spend.

Here are the five most common mistakes I see companies making:

👇 #1 - No clear marketing strategy

This might sound generic, but it's a killer. Too many accounts are run without any clear direction.

Are you running this campaign structure because you read about it on Twitter or because it solves a problem for your company?

It looks like @Meta has sent out new terms for self-serve Ads Manager accounts, and there are some huge red flags in here.

Highlights in a thread 👇

facebook.com/legal/self_ser…

S/O @jaketheadnerd for sharing this. 1⃣

This feels 'classic' Meta, but this is incredibly annoying:

"Your ads must comply with all applicable laws, regulations, and guidelines, as well as our Advertising Policies..."

Followed by:
"We may reject or remove any ad for any reason."

Which is it?

These 5 websites are a CHEAT CODE for marketers:

🧵👇 marketersplaybook.co

The other day I dropped a newsletter highlighting a 5 step process for auditing performance on your Facebook ads manager account, and how to optimize it.

Here's the summary 👇

1/24 (strap in!) Before we get started, we need to determine which metric we are optimizing for. In this case, it's ROAS. I know that ROAS is generally incorrect, but for sake of argument, we will go with this.

Yesterday I dropped a newsletter about how I added 500k+ people to my MOF retargeting audiences at a fraction of a cent each, plus a custom metric to help make better sense of it.

Here's a quick summary 👇

1/9 Things in Facebook Ads World are changing. As we get deeper into the realm of attribution loss, tried and true tactic like optimizing for purchase throughout the funnel become less efficient. That's not to say it doesn't still work, but it might be time for some adjusting.

2/9

Let’s talk Facebook Automation & Rules

I've been seeing a lot of uncertainty about how to use rules on Facebook Ads Manager lately, so I've decided to put together a list of my personal favourites. I’ve even given them cool names (a thread👇)

1/14 First, a note about rules. Rules are not only helpful for automating processes, but also for being notified of when a specific event occurs. Being notified is key because it will let you know when it’s time to take action.

2/14

Let's talk CCPA.

I had a conversation with my team at Facebook about their thoughts, and while they aren't legally allowed to tell us what to do, I was able to gather some important information for everyone. More details below:

1/12 What is CCPA:

☑️Data protection policy for residents of California
☑️All data from California is neither being collected by the pixel, nor is it being reported in your dashboard as of the hard date of July 1st

2/12

Facebook's new(ish) ABC Lookalike audiences are 🔥

Over the last 30 days, I spent ~$120,000 testing the new ABC Lookalike audiences in partnership with FB, and the performance coming out of them was solid. Results below:

1/13 First, what ABC LALs are:

☑️A lookalike audience that uses a campaign or ad set as the seed, instead of a single conversion event
☑️Uses the API to pull signals, which updates every couple of hours
☑️Does not cannibalize seed audience (little to no overlap)

2/13

I made a post the other day about my master list of wants for Facebook Ads Manager, and I think I've finished. I left out obvious bugs and opted for more core functionality/UI.

Also, if any of these already exists please let me know. FB Ads wants, a thread:

1/7 FUNCTIONALITY:

☑️Fully functional mobile app
☑️Downloadable Ads Manager desktop client
☑️Collection ad dark posts
☑️Editable thumbnail for Collection ad videos
☑️Custom formula columns (ex. contribution margin, LTV, etc)
☑️Change attribution model (a la Google Ads)

2/7

I just wrapped up a month-long, $40,000 test on Facebook Ads Manager to determine which bidding modifier worked best in top of funnel at scale. Details and results below:
(1/14) I ran three separate campaigns over a 4 week period. Each campaign ran at the same budget, but had a different bidding modifier. One was set to lowest cost, one set to cost cap, and one set to bid cap. Our main KPI was CPA, though we monitored changes to all metrics.
(2/14)