CreatorsAI 2.0 - Frank Kern

tags:: #source/video-course Frank Kern

Week 1: Winning Before You Start

Take the pressure off. Relax. No rush. It's all good.

MW: He repeats very often that he's not an expert in many things.

"What's gonna be fun and cool to sell and create? Also make sure people actually want to buy it."

We will analyze top selling courses and marketing to figure out what's already working.

Sales Copy

Good old long form sales letter is excellent for selling under $500 stuff.

This is the first prompt I used with Claude: "The text I've pasted here is a series of seven different attempts to overcome someone's objections and get them to purchase a course that shows them how to write a book in 30 days. Please read what I've pasted and let me know you understand what it is. Then I'll give you more instructions." (Then I pasted in all the close stacks that my custom bot had written.)

Once it told me it understood, I used this prompt: "OK great. Instead of presenting this as a list of "Stacks", I need you to tie them all together and separate them by using relevant sub-headlines."

It kept trying to re-write everything so I used a prompt with some examples.

This is the prompt I used: "DO NOT RE-WRITE the actual stacks themselves.

Simply tie them together with sub-headlines.

For example, stack #1 is really addressing the objection that taking this course sounds like a lot of work and is kind of expensive.

In this case, we'd simply replace the words "Stack #1: The Pros and Cons" with a sub-headline that says "Wait. Is This Really Worth It?"

Stack #2 is addressing the objection where the reader feels like they won't have enough time to complete the course.

So for stack #2, we'd replace the words "Stack #2: "What I'm Really Hearing Is..." with a sub-head that says "What If I Don't Have Time?"

...And we'd want to do similar things for each subsequent stack.

Again, we don't want to re-write the stacks themselves. We just want to tie them together with relevant subheads.

Here are the original stack descriptions:" (Then I pasted my original close stacks again).

That prompt worked pretty well.

Next, I switched back to the "KernBot" from phase 1 and I had it re-write some of the copy for each stack so it would sound more like me.

This is the prompt I used with that bot: "Please re-write this text. Don't expand on it or anything. Just re-write it:" (And then I'd paste the text for an individual stack).

This is the KernBot from phase 1:

https://poe.com/FrankBotVZ2YZUTS1W

I also just made another bot that's simply trained to take any text and re-write it using my style.

You might want to try it.

Here it is:

https://poe.com/WriteLikeFrank

You don't have to tell it to write like me because it's already trained to. You can just give it some text and say "Re-write this".

Then use the write like frank for every section.

Week 7: Advanced Facebook + Branding Ads

Two types of ads:

Frank's book funnel:

Understand your average customer value in the first 4 weeks, and compare that with the ad costs.

Branding Ads

Use Advantage+ so "feed Facebook's brain as much information as possible about our ideal customer."

Run ads for pure content to feed Facebook's data. Then use Advantage+ ads so Facebook knows which people to show your broad ads to.

SuperFame

Low budget pure content ads to specific (non ADV+) audience. Strict small match based on behaviors only. Many different pieces of content.

This helps build your brand because you look like a famous person.

This also builds the engagement data.

We also do branding to retargeting audience. \

Images get more organic views than videos.

infomarketingblog.com for inspiration

answerthepublic.com, pick a topic, get the questions and use AI to create answers for it

We want to saturate the market with the answers to these questions and drive them to our domain.

No native CTA, just a small one in the text or video post itself.

Early Ads

When you're spending less than $500/day on ads.
Meiser MEthod

Purpose:

Method:

GoHighLevel tracking parameters:

utm_source=fb_ad&utm_medium={{adset.name}}&utm_campaign={{campaign.name}}&utm_content={{ad.name}}&campaign_id={{campaign.id}}

In the example: drive people to opt-in page for on-demand class in Magic Funnel.

Select "Sales" as objective, even though it's about leads. I don't want cheap leads, I want better quality leads. Maximizing value of conversions increases price of traffic, Frank experiments between the two.

Cost per result goal - for advanced people if you really really trained it.

Good to start around 6AM in the morning your time.

Some countries love opting in for stuff and not buying anything.
Frank: US, Canada, Australia,... ("Big Four")

Audience:

Low budget + long time = it's really good at training the algorithm.

Always start with image ads because they're easy.
Turn off multi-advertiser ads because Frank doesn't know yet if it works or not.

Ad type:

Copy: something "watch this if you want to create online courses with AI." Make sure the audience knows without having to expand to read more.

Dog Whistle copy (Dan Kennedy), only our perfect audience will want to read it.

"Low budgets per ad set I think make the algorithm work harder."

I recommend you do 10 of these so $50 per day. Most are going to suck. But 1-2 are going to work. Later we'll turn off the bad ones and scale the good ones.

Speeding Up Sales

Get at least 100 sales first.
Check "Facebook Customer Hunting 101" course.

The first sale is the hardest to make. So the low ticket first sale is great if it just gets there at break even. If you do it right, it's easy to sell more stuff.

Typical lead generation in online marketing or B2B space: $5-$30 bucks.

Don't make one of these if you don't know the value of your customers.

When you have a Miser Method winner, this is how to start scaling it.

Campaign bid: cost per result goal.

International audience is usually cheaper traffic than US.

Leads from a 2-step order form even if they abandoned are great leads.