Prompt Playbook: Selling without Selling PART 5

Today we’re going to experiment with advertising. With Meta ads in particular. Why Meta? Easiest and cheapest to get running and what we learn here will help in other forms of paid media.

Everyone overcomplicates Meta ads.

They create 47 different audiences. Test 23 different interest combinations. Build complex funnel strategies. Manually optimise every setting.

I used to do this. A lot of work. And not very effective!

Meanwhile, Meta has spent billions building an algorithm that's smarter than all of us at finding buyers.

I’m running some ads right this moment and all of our hand-crafted headlines have been beaten by their suggested ones… kinda humbling.

Your job isn't to outsmart Meta. It's to give their algorithm good creative and clear goals, then get out of the way.

Today, I'm showing you the laziest, most effective way to run ads. We're going wide with targeting, feeding Meta multiple options, and letting their machine learning do what it does best - find people who buy.

$5-10 per day. Broad targeting. Multiple creative options. Let Meta figure it out.

This isn't the advanced strategy. This is the "I'm new and want to learn what works" strategy. And honestly? It often beats the "expert" approach…

Let's get started

Fair warning. This is going to read like a checklist. Cause…it sort of is! I want to give you an overview of what to do and the best way is just listing it all out. Work alongside in the Meta ads interface for best results.

Also…if (and when!) you get stuck go ahead and ask AI. AI can give you a complete ads integration plan no problem.

Phase 1: The Technical Setup

First, the boring but essential bits

Step 1: Create Meta Business Account

• Go to business.facebook.com

• Create account (just need business name)

• Add your Facebook page

• Add payment method

Fair warning the business.facebook.com interface is trash. Get ready for frustration. Everyone hates it. Without exception!

Also, turn off you AdBlock on the Meta ads interface pages. It can muck up the interface.

Step 2: Get Your Pixel Code

Your pixel code is a little snippet of code that’ll connect your website and product to Meta’s ad interface. Basically it allows tracking so you can tell who signs up and buys from your adverts.

• In Events Manager, create a pixel

• Name it something simple

• Copy the pixel ID (looks like: 1234567890)

Step 3: Tell Lovable to Set It Up

Adding the code can be a pain. But we’ll use Lovable.

Here's the prompt for Lovable

Please add the Meta Pixel to my site with these events:

Pixel ID: [YOUR PIXEL ID]

Events to track:
1. PageView - all pages
2. Lead - when someone enters their email
3. InitiateCheckout - when someone clicks buy/subscribe
4. Purchase - on payment confirmation

Add the base pixel code to the head of all pages, then add the specific event tracking to the relevant actions.

Make sure it's working by showing me where you've added the tracking.

Lovable will handle the technical implementation. You just paste and wait.

Once it’s done visit your page in an anonymous browser to “trigger” a visit. Then check in the Meta ad interface and your pixel should now show as active. If still not then the visit didn’t get recorded and something is up - work with Lovable to get it sorted.

Phase 2: Gather Creative

Your ad creative will be your images/videos as well as text like headlines and description text.

We are using our short videos as the primary visual creative. For the text we want 5 variations of each element. Why 5? Meta can run 5 against each other and optimise for the winner.

Here’s a prompt to generate

I need Meta ad copy variations for testing. Create 5 versions of each element:

My product: [Your product]
My VSL hook: [Your main message]
Target customer: [Who they are]
Main benefit: [What they get]
Price: [Your price]

Generate:

5 HEADLINES (max 40 characters each)
- Focus on different angles
- Test benefits vs curiosity vs problem
- Keep them punchy and specific

5 PRIMARY TEXTS (max 125 characters each)
- The main ad copy above the creative
- Test different opening hooks
- Include different emotional triggers
- Mix problem/solution/benefit angles

5 DESCRIPTIONS (max 30 characters each)
- The link description text
- Clear value statements
- Different urgency levels

All copy should:
- Speak to ONE person not crowds
- Use simple, conversational language
- Focus on transformation not features
- Have clear but not pushy CTAs

This gives Meta options to find winners.

We’ll use these in campaign setup. For now just hold on to them. And make sure you have video files of your short sales videos ready.

Phase 3: Campaign Setup

Here's where most people go wrong - they optimise for clicks or traffic. Who cares about visits. We want BUYERS, so we optimise for conversions from day one:

Campaign Level

• Objective: Sales (yes, even with small budget)

• Campaign name: "[Product] - Broad Test - [Date]"

• Advantage campaign budget: ON (let Meta distribute)

• Budget: £7/day to start

Why sales objective? Meta's algorithm learns who converts, not just who clicks. Even with few conversions, it starts learning buyer patterns immediately.

Alternatively you can optimise for Leads (ie. sign ups). This will depend on your product and business honestly.

Phase 4: Audience Plus - Going Wide on Purpose

This is counterintuitive but crucial. Instead of telling Meta who to target we’ll let Meta work it out. Basically we go wiiiiiide and let Meta work it out.

The main variable we’ll choose in country. I tend to target US, Canada, UK, Australia, NZ as primary. Your market may differ based on your product of course!

Ad Set Settings

• Name: "Broad - Advantage Plus"

• Conversion event: Purchase

• Audience: Advantage+ Audience (Meta's auto-targeting)

• Location: Your countries (UK, US, etc.)

• Age: 25-65 (yes, that wide)

• Gender: All

• Languages: Your market's language

• Detailed targeting: SKIP IT

Why go broad?

• You don't know your buyers yet!

• Meta's algorithm knows more than you

• Narrow targeting often misses unexpected buyers

• The algorithm needs room to learn

• Your creative will naturally filter

What we’re doing here is not just trying to make sales. It’s also seeing who responds to what we have to offer. We’re doing research at the same time as we sell!

Phase 5: Creative Plus - More Options, More Learning

We use a similar method with creative. We don’t hand craft creative campaigns, carefully matching headlines to videos. Nope. We chuck it all in the pot and let Meta cycle between visual assets, headlines, descriptions and more until it finds the best mix.

Because, frankly, it generally does a better job!

Upload ALL your creative options

Creative Setup

• Format: Flexible (let Meta choose)

• Turn on Advantage+ creative optimisations

• Upload all 3 video versions from yesterday

• Add all 5 headlines

• Add all 5 primary texts

• Add all 5 descriptions

• Call to action: Learn More

This creates 3 x 5 x 5 x 5 = 375 possible combinations. Meta will test and find winners automatically. You don't manage 375 ads. Meta does. You just see which combinations win as we begin testing.

Phase 6: Launch and Learn

OK once we’ve got everything loaded up we can turn on the campaigns.

Your launch checklist

1. Set £7-10 daily budget

2. Set £50 total spending limit (safety)

3. Turn on campaign

4. LEAVE IT ALONE for 72 hours minimum

Why 72 hours?

• Day 1: Meta is spraying everywhere

• Day 2: Starting to learn patterns

• Day 3: Beginning optimisation

• Day 4+: Actually optimising properly

Touching it early breaks the learning. You’ll see next to the campaign it’ll say “Learning”. During that period if we stop the campaign we don’t really get any useful information - so let it do its thing.

Reading the Data (Even Without Sales)

OK so a few days pass…now what? Well hopefully you’ve seen some sales activity. If so we’re rocking and a rolling. We look at how much it cost us to get a customer and we scale ads accordingly.

If it cost us £10 for a £30 sale then … spend more on advertising! We can scale up the ad budgets and get rolling. We’re making money.

Most likely you won’t have seen a sale. Don’t worry. The above scenario is rare (but it does happen!).

The good news is that after 72 hours with ~£20-30 spent, you'll have data even if you made zero sales.

We now know a whole bunch of things about our marketing. Some examples + how to get the data follows:

What Headlines Worked?

• Check Breakdown → By Dynamic Creative Asset

• See which headlines got highest CTR

• These words resonate with your market

• Drop other headlines and write more that are close to the working headline.

Which Videos Performed?

• Same breakdown shows video performance

• Lowest cost per click = most engaging

• Make more like the winner, less like the losers.

Who's Actually Clicking?

• Breakdown → By Age, By Gender, By Geography

• Surprises are common ("My audience is women 45-54, not men 25-34!")

• This informs everything going forward, especially in messaging and targeting

Etc etc etc. We can get a tonne of information from running ads. And this intel is genuinely more valuable than the sale or two you may have made. It’s worth thousands if we listen to the data.

From here we basically iterate. Keep your budgets low, no need to go wild yet. In Week 2 do more creatives and headlines like those that worked best. The next week maybe narrow down your targeting. Then as you start to see traction nudge budgets up slowly.

Just make sure you keep learning!

Your Deliverable Today

Lots to do today! Take your time on this one as it can be fiddly!

1. Set up Meta Business Manager

2. Get pixel installed via Lovable

3. Generate text variations with prompt

4. Create campaign with Advantage+

5. Upload creative and launch

6. Set calendar reminder for 72 hours

7. Do NOT touch until reminder goes off

Once we’ve done this core setup thankfully Meta does all the heavy lifting for us.

Build in Public Component

Share your launch

"Day 45 of AI Summer Camp: Meta ads live!

Strategy: Going broad, letting algorithm learn Budget: £7/day Creative: 3 videos, 375 text combinations! Expectation: sales would be nice but right now this is a learning budget.

I’m going to assume the algorithm knows more than me. Time to let it teach me who my buyers actually are."

What's Next?

Week 9 complete! We’ve come all the way from working out your personal blocks to selling to actually having a basic advertising sales system up and running for us.

Sales might not come immediately. That's not failure - it's education. Every click tells you something. Every non-click tells you more. So keep budgets manageable and keep learning. Kinda the watchword of this whole AI Summer Camp honestly!

Keep Prompting,

Kyle