Solved: CBO Campaign Budget Heavily Skewed to One Ad Set

Hi there,

Thanks for reaching out about the issue with your CBO campaign. It’s a classic problem that trips a lot of people up, but the solution isn't quite as simple as just pausing the ad set that's hogging all the budget. The fact that Meta is piling the budget into one place tells us something really important about your campaign, and fixing it is more about your overall structure than just flicking a switch.

We'll need to look at why CBO is doing this...

Okay, so first things first, let's get one myth out of the way. Campaign Budget Optimisation (CBO) isn't being "unfair" or "broken" when it spends 99% of your budget on one ad set. It's actually being brutally, coldly efficient. You've given it a pile of cash and one simple instruction: "get me the best results possible for this money". And that's exactly what its trying to do.

When you launched your campaign, the first ad set got approved and started running. Instantly, it began generating data for Meta's algorithm. It got impressions, maybe some clicks, maybe a low-cost "add to cart". The other ad sets, stuck in review, were providing zero data. From the algorithm's perspective, it had one option that was working (even if just a little) and three options that were doing absolutely nothing. So, it did the logical thing: it pushed all the budget to the only horse in the race.

It doesn't care about giving each ad set a "fair chance". It's not a parent trying to treat all its children equally. It's a machine programmed for one thing: efficiency. I've seen this happen countless times. A client comes to us wondering why CBO "hates" their new audiences, but the reality is the campaign structure set them up to fail from the start. You're essentialy asking the algorithm to compare an audience with some performance history against audiences with none. The outcome is pretty much decided before the race even begins. This is why procedural stuff, like ensuring all your ad sets are ready and approved to launch simultaneously, is so important. A small oversight at the setup stage can completely invalidate your test and waste your budget.

I'd say you shouldn't pause it... just yet.

So your first instinct is to pause the ad set that’s eating the budget to "give the others a chance". I get it, it feels like the right thing to do. But you need to be careful. If that ad set is, as you say, getting "somewhat good return", pausing it could do more harm than good. You'd be stopping an ad set that's potentially making you money and forcing the budget onto unproven audiences. Even worse, you'll be knocking that ad set out of the learning phase, and when you turn it back on, it might not perform as well.

Instead of hitting pause, here's what I'd do in your exact situation:

1. Wait at least 72 hours. Seriously. Don't touch anything. Meta's algorithm needs time to adjust. Now that the other ad sets are approved, it will start to slowly feed them a little bit of budget to see how they perform. If they start delivering results at a competitive cost, you'll see the budget distribution start to even out naturally. If they don't, Meta will continue to favour the original ad set. Patience is a massively underrated skill in paid advertising.

2. Consider a small budget increase. If after a couple of days the spend is still massively skewed, try increasing the total CBO budget by about 20%. This sometimes gives the algorithm a bit more "breathing room" to experiment with the other ad sets without having to take too much away from the one it already knows is working. It's not a guaranteed fix, but it's a much less disruptive option than pausing.

3. Analyse the "winner" properly. You said it's getting a "somewhat good return". We need to be way more specific than that. What's the Cost Per Purchase? The ROAS? Is it actually profitable based on your product margins? If this ad set has a ROAS of 4x and the others are struggling to get any conversions, then CBO is doing its job perfectly. Maybe the other ad sets just aren't as good. You can't blame the tool for telling you something you might not want to hear about your audiences.

To help you visualise this, here's a little simulator. Play around with the potential performance of your ad sets. See how a CBO-like logic would distribute the budget based on which audience is actually performing, not which one you *want* to perform. Notice how even a small difference in performance can lead to a big shift in spend over time.

You probably should rethink your campaign structure for testing.

This is the real heart of the issue. The budget problem you're seeing is just a symptom of a flawed testing structure. To get reliable results and scale your account effectively, you need a much more deliberate and organised approach. The goal of a testing campaign is to find winning audiences, and you can only do that with a fair test.

The biggest mistake I see people make is lumping completely different types of audiences into one CBO campaign. They'll have an ad set with a cold interest audience, one with a 1% Lookalike of purchasers, and another retargeting website visitors from the last 30 days. This is a recipe for disaster. The retargeting audience will almost always win because it's a 'hot' audience of people who already know you. CBO will pour the budget into it, and you'll learn absolutely nothing about whether your cold audiences could have worked. It's like putting a Formula 1 car in a race against a couple of family hatchbacks and then being surprised when it wins.

The solution is to structure your campaigns based on the marketing funnel: Top of Funnel (ToFu), Middle of Funnel (MoFu), and Bottom of Funnel (BoFu).

• ToFu (Top of Funnel): This is your prospecting campaign. It's for reaching people who have never heard of you before. All your cold audiences go in here: interest-based targeting, broad targeting, and lookalikes based on customer data.
• MoFu (Middle of Funnel): This is for people who have engaged with you but haven't taken a key action yet. Think video viewers, social media page engagers, people who visited your blog but not a product page.
• BoFu (Bottom of Funnel): This is your hot retargeting audience. People who have visited product pages, added to cart, initiated checkout, etc. These are the people closest to buying.

By splitting your campaigns like this, you ensure CBO is only ever comparing like with like. In your ToFu campaign, CBO will work to find the most effective *cold* audience. In your BoFu campaign, it'll find the most effective *hot* audience. This gives you clean data and allows you to make much smarter decisions about which audiences to scale and which to turn off. It's how we structure accounts for all our clients, from eCommerce stores to B2B SaaS companies. I remember one eCommerce client selling women's apparel who saw their return jump to over 690% just by restructuring their chaotic account into this simple ToFu/MoFu/BoFu framework. It works because it's logical.

You'll need to define what 'good return' actually means.

This is probably the most important shift in mindset you need to make. You can't optimise what you don't measure, and "somewhat good return" isn't a metric. To make smart decisions, you need to know exactly how much you can afford to spend to acquire a customer. This is where most business owners get it wrong. They focus obsessively on lowering their Cost Per Lead (CPL) or Cost Per Click (CPC), without realising that the cheapest leads are often the worst quality.

The key metric you need to understand is Customer Lifetime Value (LTV). How much profit does a customer generate for your business over their entire relationship with you? Once you know this, you can work backwards to determine your maximum allowable Customer Acquisition Cost (CAC).

A healthy business model typically aims for an LTV:CAC ratio of at least 3:1. This means for every £1 you spend to acquire a customer, you get £3 back in lifetime profit. Let's do the maths.

Average Revenue Per Account (ARPA): How much a customer spends on average per month. Let's say £100.
Gross Margin %: Your profit margin on that revenue. Let's say 70%.
Monthly Churn Rate: The percentage of customers you lose each month. Let's say 5%.

The calculation is: LTV = (ARPA * Gross Margin %) / Monthly Churn Rate

So, in this example: LTV = (£100 * 0.70) / 0.05 = £70 / 0.05 = £1,400.

Each customer is worth £1,400 in gross margin. With a 3:1 ratio, you can afford to spend up to £466 (£1400 / 3) to acquire a single customer and still have a very healthy business. Suddenly, a Cost Per Purchase of £50, £100, or even £200 doesn't look so bad, does it? It looks like a bargain. This is the maths that lets you scale aggressively. Without it, you're flying blind, turning off ad sets that might actually be goldmines because they "feel" expensive.

Use this calculator to figure out your own LTV and target CAC. Be honest with your numbers. This single calculation is more valuable than any "Facebook hack" you'll ever find.

I've detailed my main recommendations for you below:

To pull all this together, here’s a clear, actionable plan based on what we’ve discussed. This is the kind of strategic framework we implement for clients to move them from chaotic, unpredictable results to a structured system for growth.

I know this is a lot to take in, and it's a world away from just "pausing an ad set". But this strategic depth is what separates accounts that spend a bit of money and get some results from accounts that truly scale and build a predictable engine for growth. It involves understanding the mechanics of the platform, the economics of your business, and the psychology of your customer.

Getting this all set up correctly and managing it day-to-day can be a full-time job. It's why businesses often turn to specialists. Instead of you having to become an expert in CBO algorithms, audience testing frameworks, and LTV calculations, you can have someone manage the entire process for you, applying proven structures and strategies from day one.

If you'd like to have a chat about how this kind of strategic approach could be applied to your business, we offer a free, no-obligation initial consultation. We can go through your ad account together, look at your current setup, and give you some more specific, actionable advice. It's often a really helpful way to get a taste of the expertise you'd be getting if we were to work together.

Hope this has been helpful and gives you a clearer path forward.

Regards,

Team @ Lukas Holschuh