Solved: Increasing Facebook Ads Budget During Learning Phase?

Hi there,

Thanks for reaching out!

Happy to give you some of my initial thoughts on your question. It's a common one, and it's great you've found an ad set that's working well with a 3.3 ROAS. That's a solid start. Most people in your position think the next step is to just carefully nudge the budget up, but honestly, you're probably focusing on the wrong thing if you really want to scale this thing.

The real path to scaling isn't about tiny budget increases, it's about building a robust system that can handle much larger spend without breaking. Let's get into it.

You're asking the wrong question about the Learning Phase...

First, to answer your direct question: is there any harm in increasing the budget by 10-15%? Probably not. A small increase like that is unlikely to completely reset the learning phase or throw the algorithm into a spin. It's the cautious, textbook move. But it's also a massive distraction.

The obsession with staying out of the "Learning Limited" status is a trap. It makes advertisers timid. It makes them scared to make meaningful changes that could actually lead to scale. The truth is, almost every significant action—a new creative, a new audience, a major budget shift—will cause some temporary re-learning. That's not a bug; it's how the machine works. It needs to process the new information. Your goal shouldn't be to avoid the learning phase at all costs, but to give the algorithm enough high-quality data and options that it learns quickly and finds you more customers.

The real risk you're facing isn't resetting learning on one ad set. The real risk is that this single ad set will eventually fatigue, your ROAS will drop, and you'll have nothing else ready to take its place. You're trying to protect a single golden goose when you should be building a farm. Scaling isn't about feeding one goose a bit more grain each day. It's about breeding more geese.

Relying on one ad set, no matter how good it is right now, is a huge bottleneck. To go from $150/day to $500/day, or $1,000/day and beyond, you need a proper strategy for audience expansion. That's your real job now.

We'll need to look at your audience strategy...

Your current success is brilliant, but it's on a very narrow foundation. You've found one pocket of customers that likes your offer. To scale, you need to find more pockets. I've audited hundreds of ad accounts, and the ones that scale successfully all have one thing in common: a structured approach to audience testing. They dont just get lucky with one ad set.

We usually prioritise audiences based on how close they are to making a purchase. The further down the funnel they are, the better they tend to perform. For an eCommerce account like yours, the hierarchy looks something like this. You should be building and testing these in a structured way.

META (Facebook/Instagram) ADS AUDIENCE PRIORITISATION

BoFu (Bottom of Funnel - Hottest Audiences):

These are people who have shown strong buying intent. They are your lowest hanging fruit for retargeting.

• -> Added to cart (last 7-14 days)
• -> Initiated checkout (last 7-14 days)
• -> Previous purchasers (last 180 days)
• -> Highest value previous customers (create a custom audience from your customer list)

MoFu (Middle of Funnel - Warm Audiences):

These people are aware of you and have shown some interest, but aren't quite ready to buy. They need a bit more persuasion.

• -> All website visitors (last 30 days)
• -> Viewed specific product pages (last 30 days)
• -> Engaged with your Facebook or Instagram page (last 90 days)
• -> Watched 50% of your video ads (last 90 days)

ToFu (Top of Funnel - Cold Audiences):

This is where true scale comes from. These are new people who haven't heard of you before. Your current winning ad set is a ToFu audience.

• -> Lookalike Audiences (LALs). These are your most powerful tool for finding new customers. You should create them based on your best data sources. The priority would be:
  1. LAL of your highest value customers list
  2. LAL of your full purchasers list
  3. LAL of people who initiated checkout
  4. LAL of people who added to cart
• -> Detailed Targeting (Interests, Behaviours). This is what you're doing now. The trick is to get specific. If you sell high-end coffee beans, dont just target "Coffee". That's far too broad. You'd target interests like "James Hoffmann", "Fellow Products", "Aeropress", or competitor brands like "Blue Bottle Coffee". You need to think what specific things your ideal customer is interested in, not what the general population likes.

You have 38 purchases. That's almost enough to start building some powerful lookalike audiences. You need a source audience of at least 100 people for a lookalike to work, but the more data you have, the better. As soon as you cross that 100-purchase threshold, your number one priority should be creating a 1% Lookalike of your purchasers and testing it in a new ad set. This is your next step to scale, not a 10% budget increase.

I'd say you need a proper campaign structure for this...

A single ad set in a CBO campaign is fine for an initial test, but it's not a scalable structure. The whole point of Campaign Budget Optimisation (CBO) is to give the algorithm a budget and let it automatically distribute it across multiple ad sets, favouring the ones that perform best. By only giving it one ad set, you're not using the tool as its designed.

A much better, more resiliant structure is to seperate your campaigns by funnel stage. This allows you to control your messaging and budget allocation more effectively. Here’s a simple but powerful structure that we use for many of our eCommerence clients.

With this setup, you can launch your new prospecting ad sets (like the lookalikes) inside the Prospecting CBO. The algorithm will start feeding them budget, testing them against your current winner. If a new ad set performs better, CBO will automatically shift more spend to it. If it performs worse, it'll get less spend. This is how you scale safely and systematically. You're not gambling by turning off a winner; you're creating a competitive environment where new challengers can prove their worth.

You can start with a modest budget on the Retargeting campaign. The Prospecting campaign will do the heavy lifting of bringing in new people, and the Retargeting campaign will efficiently convert those who showed interest but didn't buy on their first visit. This structure is what allows our clients to scale spend while maintaining a healthy ROAS. I remember one women's apparel client for whom we generated a 691% return by implementing a very similar structure, focusing heavily on getting the retargeting right.

You probably should calculate what you can actually afford to spend...

A 3.3 ROAS sounds good. But what does it actually mean for your business? Is it good, great, or just okay? The only way to know is to understand what a customer is actually worth to you over their lifetime. This is probably the most overlooked metric by advertisers, but its the one that separates the amateurs from the pros.

The real question isn't "How low can my Cost Per Acquisition (CPA) go?" but "How high a CPA can I afford to acquire a truly great customer?" The answer is in your Lifetime Value (LTV).

Let's run through a hypothetical calculation for an eCommerce store. You'll need to plug in your own numbers, of course.

• Average Order Value (AOV): What's the average value of a single purchase? Let's say it's £60.
• Purchase Frequency: How many times does a customer buy from you in a year? Let's say it's 2.5 times.
• Customer Lifetime: How many years does a customer typically stay with you? Let's say it's 2 years.
• Gross Margin %: What's your profit margin after cost of goods? Let's say it's 70%.

The calculation for the profit you make from a customer over their lifetime would be:

LTV = (AOV * Purchase Frequency * Customer Lifetime) * Gross Margin %

Let's plug in our numbers:

LTV = (£60 * 2.5 * 2) * 0.70

LTV = (£300) * 0.70 = £210

This is the truth. In this example, each new customer you acquire is worth £210 in gross profit to your business. Now let's look at your current ad perfomance.

Your ROAS is 3.3. Your CPA is your spend divided by your purchases. With a $150/day budget and 38 purchases in 3 days, your total spend was $450. Your CPA is $450 / 38 = $11.84 (or about £9.50).

You are paying £9.50 to acquire a customer that will generate £210 in profit for you. That's a LTV:CAC ratio of 22:1. That is incredible. A healthy ratio is typically considered to be 3:1. You have an enormous amount of room to be more aggressive.

Knowing this should completely change your mindset. You're not trying to protect a 3.3 ROAS. You're trying to acquire as many customers as possible for a CPA that is profitably below your £210 LTV. You could let your ROAS drop to 2.0, or even 1.5, and still be printing money. Understanding this math is what gives you the confidence to scale aggressively. One campaign we worked on for a SaaS client involved reducing their £100 CPA to £7, which was only possible because we first understood their LTV allowed for much higher initial acquisition costs.

This is the main advice I have for you:

To pull this all together, here is a summary of what I believe your focus should be. Forget the 10% budget increases for now. That's playing not to lose. It's time to play to win.

I know this is a lot to take in. It's a shift from just 'running ads' to building a proper, professional advertising system. It's the difference between a campaign that gets lucky for a few weeks and a business that can scale profitably for years. This process of research, building, testing, analysing, and optimising is what we do all day, every day for our clients.

This is how we've achieved results like a 1000% ROAS for a subscription box client or helped a B2B SaaS company generate over 1500 trials. It's not about a magic trick; it's about applying a rigorous, proven process.

If you'd like to chat through this in more detail for your specific business and have us take a proper look under the hood of your ad account, we offer a free, no-obligation 20-minute strategy session. It might help you get a clearer picture of the path forward.

Either way, hope this was helpful and gives you a new way to think about scaling.

Regards,

Team @ Lukas Holschuh