Solved: Stuck in Facebook Ads Learning Phase - Need Help!

Hi there,

Thanks for reaching out! I saw your query and I'm happy to give you some initial thoughts and guidance based on my experience. It sounds like you've got a decent start with a profitable CPA, which is better than where a lot of people are when they first start out. But getting stuck in the 'learning phase' is a common headache, so let's break it down.

What you're asking is a really good question, and your thinking isn't dumb at all. It shows you're paying attention to the mechanics of the platform. But just cranking up the budget isn't always the simple fix it seems to be. I'll walk you through my thoughts on the whole situation, from the learning phase itself to the stuff that actually makes a difference in the long run.

So, what about this 'learning phase' then?

Right, lets get this out of the way first. The learning phase is basically Meta's algorithm trying to figure out who is most likely to buy your product. To do this, it needs data. The official line is it needs about 50 conversions (purchases, in your case) within a 7-day window to get a proper feel for things. Once it has that data, it can 'exit' learning and performance usually becomes more stable and often, a bit cheaper. Your problem is that with a $140 CPA and a $200/day budget, you're getting maybe 1-2 sales a day, which means you're nowhere near the 50 sales a week needed.

So your logic to increase the budget to $1000/day to hit that 50-conversion target is sound, in theory. If you could maintain that $140 CPA, you'd get roughly 7 sales a day, hitting the 50 mark in a week. The big 'if' here is whether the CPA will actually stay the same. When you suddenly hike the budget by 5x, you force the algorithm to go find a lot more people, a lot faster. This can sometimes make the CPA spike, at least for a bit, as it explores less 'optimal' pockets of your audience. You could end up just spending a load more cash without getting the stable results you want.

Honestly, the learning phase is important, but being obsessed with exiting it can sometimes be a distraction from the real problems. A campaign being stuck in 'Learning Limited' isn't the end of the world if it's profitable. The real goal isn't just to exit learning; it's to build a campaign so efficient that it exits learning naturally because it's getting loads of conversions at a good price. The root cause here isn't your budget, it's more likely your targeting and campaign setup which is leading to a CPA that makes scaling difficult. Fix the underlying issues, and the learning phase problem often solves itself.

We'll need to look at your audiences...

This is probably the most important bit. Getting the targeting right is everything. You're using an Advantage+ campaign, which gives a lot of control to Meta. That's fine, but it relies heavily on having a well-seasoned pixel with lots of data, which you don't have yet as you've only been running ads for a few days. You might be better off with a manual campaign where you can control and test audiences properly first.

I always tell clients to think about the 'funnel'. You have different types of audiences, and you should treat them differently.

-> Top of Funnel (ToFu) - People who don't know you: This is your cold traffic. You're running a new account, so this is where you need to start. Your best bet here is 'detailed targeting'. This means targeting interests, behaviours, and demographics. You need to get really specific. Who is your ideal customer for a $500 product? What brands do they like? What magazines do they read? What influencers do they follow? What software do they use? You need to build a picture of them and find interests that only they would have.

For example, I've seen people trying to target eCommerce store owners by using the interest "Amazon". That's way too broad. It includes millions of shoppers. A much better aproach is to target interests like "Shopify", "WooCommerce", or pages of popular eCom-related software. It's about finding the interests that have a high concentration of your target audience.

-> Middle of Funnel (MoFu) - People who've shown some interest: These are people who have visited your website, watched one of your videos, or engaged with your ads. You need to be retargeting them. They already know you, so they are much more likely to convert. You can set up custom audiences for website visitors, video viewers, etc.

-> Bottom of Funnel (BoFu) - People who were close to buying: These are the gold dust. People who added your product to their cart but didn't buy, or people who started the checkout process. You should be hitting these people with specific ads to get them back. Things like "Forgot something?" or offering a small incentive if needed.

For a new account, you need to focus on ToFu to gather data. I'd set up a campaign with multiple ad sets, each testing a different theme of detailed targeting interests. One ad set for competitor brands, one for related magazines/blogs, one for lifestyle interests, and so on. Run them for a few days and see which one gets the best CPA. Once you find a winner, you can put more budget behind it.

I'd say you need a better campaign structure...

Following on from the audience point, relying on a single Advantage+ campaign is a bit risky when you're starting out. It's a black box. You don't get much data on what's working and what isn't. I'd recomend a more traditional campaign structure to begin with, so you can learn and optimise properly.

A structure I often use for clients, especially in eCommerce, looks something like this:

Campaign 1: Cold Traffic / Prospecting (ToFu)
This is where you go out and find new customers. The goal here is to get people to your website. Inside this campaign, you'd have several ad sets.

• -> Ad Set 1: Detailed Targeting (Interest Group A - e.g., competitor brands)
• -> Ad Set 2: Detailed Targeting (Interest Group B - e.g., related hobbies/lifestyles)
• -> Ad Set 3: Detailed Targeting (Interest Group C - e.g., demographic-based)

Once you have enough data (you need at least 100 purchasers to make a good lookalike audience), you can add ad sets for Lookalike audiences. You'd start with a lookalike of your purchasers, as thats the highest-quality seed audience you have.

Campaign 2: Retargeting (MoFu/BoFu)
This campaign is purely to bring back people who've already interacted with you. It's usually much cheaper to convert these people.

• -> Ad Set 1: Website Visitors (last 30 days, excluding purchasers)
• -> Ad Set 2: Video Viewers (people who watched 50% of your ad, last 30 days)
• -> Ad Set 3: Add to Cart / Initiated Checkout (last 7 days, excluding purchasers) - This is your hottest audience.

By spliting your campaigns like this, you can control the budget for each part of the funnel. You can also tailor your ad creatives and messaging. The ads you show to someone who has never heard of you should be different from the ads you show to someone who abandoned their cart yesterday.

I remember one campaign we ran for an outdoor equipment brand; getting the structure right was what allowed us to properly test and scale, eventually driving over 18,000 website visitors from the winning ad sets. It gives you control, and control is what you need to bring that CPA down.

You probably should test more creatives...

You mentioned you have 3 creatives. That's a good start, but for a high-ticket item, you really need to be testing a wider variety of angles and formats. Your creative is what does the selling. Poor creatives mean a low click-through rate (CTR), which means a high cost-per-click (CPC), which ultimately leads to a high CPA.

Think about the different reasons someone would buy your $500 product. You need to create ads that speak to each of these motivations.

Here are some ideas for creative testing:

• Formats: Don't just stick to images. Test single videos, carousel ads showing different features or use cases, and even simple text-based ads in some placements. We've had several SaaS clients see fantastic results with User-Generated Content (UGC) style videos. They look authentic and build trust, which is massively important for a high-ticket sale.
• Angles/Hooks: Your ads need to grab attention in the first 3 seconds. Try different opening hooks.
  • -> Problem/Solution: Start with the problem your product solves. "Tired of X? Our product does Y."
  • -> Benefit-driven: Focus on the end result. "Imagine achieving Z with our product."
  • -> Feature showcase: A video or carousel that demonstrates the product's unique features.
  • -> Social Proof/Testimonial: Use a customer quote or a clip of a customer review. This is super powerful for expensive items.
• Copy: Test short-form copy vs long-form copy. Sometimes for an expensive, considered purchase, longer copy that tells a story and answers all the potential questions can work wonders. Other times, a short, punchy message is all you need. You won't know until you test.

You should aim to have at least 3-5 completely different creatives in each ad set. Let them run, and after a few days, Meta will start to show you which ones are performing best. Double down on the winners and turn off the losers. This constant process of creative testing and iteration is fundemental to lowering your CPA and finding a message that really resonates.

You'll need to think about what a good CPA actually is...

This is an interesting point. You said your CPA is $140 on a $500 product with a 70% margin. Your cost of goods is 30% of $500, which is $150. So your profit per sale before ad costs is $350. With a $140 CPA, your net profit is $210 per sale. That's a Return on Ad Spend (ROAS) of $500 / $140 = 3.57x. Frankly, that's very healthy!

The problem is that it's not low enough to let you scale quickly and get out of the learning phase. So, while it's profitable, it could almost certainly be better. Based on my experience running hundreds of campaigns, I can give you some benchmarks. For an eCommerce product sold in developed countries, here's what we often see:

As you can see, your current $140 CPA is quite a bit higher than the typical 'high end' we'd expect for an optimised campaign. This tells me there is significant room for improvment. If you could get your CPA down to, say, $70, you'd be getting twice as many sales for the same budget. At your current $200/day spend, that would mean almost 3 sales a day, which would get you to about 20 sales a week - still in 'learning limited', but much closer. If you then increased the budget, you'd exit learning much more easily and profitably.

So, the goal shouldn't be "how do I spend more to exit learning?", it should be "how do I make my campaign more efficient to lower my CPA?". If you can do that, the scaling and the learning phase stuff becomes easy.

This is the main advice I have for you:

To pull all this together, here’s a summary of what I'd recommend you focus on. It's about building a solid foundation first, rather than just throwing more money at the problem.

I know this is a lot to take in. It's not just about flicking a switch; it's about setting up a proper system of testing and optimisation. It's about understanding your audience, refining your targeting, creating ads that actually persuade people, and building a campaign structure that lets you see what's going on.

That's where professional help can make a huge difference. With years of experience and a deep understanding of these platforms, an expert can help you skip a lot of the expensive trial and error and get straight to the strategies that work. We've helped eCommerce clients achieve a 1000% return on ad spend by implementing these kinds of structured, data-driven aproaches.

If you'd like to go over your account in more detail and get a proper strategy mapped out, we offer a free initial consultation. We could have a look at your setup together and figure out the best path forward for you.

Regards,

Team @ Lukas Holschuh