Solved: CBO Campaign - Should I delete ad set with 0 sales?

Hi there,

Thanks for reaching out! I saw your question about your CBO campaign and thought I'd give you some more detailed thoughts. It's a really common situation to be in when you're starting out, seeing some ad sets work while others lag behind, and not being sure what to do next.

The short answer is you're right to question that non-performing adset, but just deleting it might be a bit premature and could cause you to miss out on valuable data. The real solution is a bit more nuanced and involves understanding how CBO actually works and building a more robust testing structure around it. I'll walk you through the exact framework we use, from audience selection to the metrics that actually matter, so you can stop guessing and start scaling predictably.

We'll need to look at how CBO really works...

First off, it's good you're using CBO (Campaign Budget Optimisation). It's the right way to go. The whole point of CBO is to let Meta's algorithm do the heavy lifting. You set the budget at the campaign level, and it automatically allocates more of that budget to the ad sets (and ads) that are getting you the best results for the lowest cost.

So, when you see one adset with 0 sales and the others getting all the love, the algorithm is already doing its job. It has identified that adset as a poor performer (for now) and is deliberately not spending much money on it. This is a good thing! It’s protecting your budget from being wasted.

This brings me to the first rule: don't be too quick on the trigger finger. Three days is not alot of time in the world of Facebook ads, especially if the campaign is still in the 'learning phase'. The algorithm needs data and time to figure things out. Turning off an adset too early is a classic mistake. You might kill a potential winner that just needed a bit more time or budget to find its groove.

So, what should you do? Instead of deleting it, you should just turn it off. Deleting erases all the data associated with that adset, which is valuable for future analysis. Turning it off preserves that data. But the bigger question is *when* to turn it off. My rule of thumb is to have a clear 'kill metric'. Usually, this is based on your target Cost Per Acquisition (CPA) or cost per sale. If you know you're willing to pay, say, £20 for a sale, you should let a new adset spend at least 1x to 2x that amount (£20-£40) before you make a decision. If it has spent £40 and hasn't got a single sale, while your other adsets are getting sales for £15, then you can confidently turn it off. It's had its chance and failed the test.

Without a clear metric like this, you're just making decisions based on feelings, and that's a surefire way to lose money.

I'd say you need a better audience strategy...

Your current problem with one underperforming ad set is really a symptom of a larger issue: your overall audience and campaign structure. Just having five ad sets in one CBO is a start, but it's probably not structured for long-term success and scaling. To get consistant results, you need to think in terms of a funnel.

I usually structure accounts into three core campaign types, mapping to different stages of the customer journey:

• Top of Funnel (ToFu) - Prospecting: This is for finding new customers. Your ad sets here will target cold audiences based on interests, behaviours, and eventually, lookalikes. This is where your five ad sets currently live.
• Middle of Funnel (MoFu) - Re-engagement: This is for people who've shown some interest but haven't taken a high-intent action. Think website visitors, video viewers, or social media engagers. You're warming them up.
• Bottom of Funnel (BoFu) - Retargeting: This is for people who are close to buying. They've added a product to their cart or started the checkout process. These are your hottest leads, and you need to get back in front of them with a compelling reason to finish their purchase.

Running these as separate, always-on campaigns (each with their own budget) is far more effective than lumping everything together. It allows you to tailor your messaging and offers to where the user is in their journey. You wouldn't speak to a total stranger the same way you'd speak to someone who's about to hand you their credit card, would you? Your ads should be no different.

Here's a visual breakdown of how that structure looks and the audiences you should prioritise testing within it.

When you're starting out, most of your effort and budget will be in the ToFu campaign. The key is to be smart about the interests you test. Don't just pick broad interests like "Shopping". Think about your ideal customer. What magazines do they read? What brands do they love? What influencers do they follow? What software do they use? Targeting pages or interests related to these will give you a much higher quality audience. For an eCommerce store, this might mean targeting competitor brands, relevant lifestyle magazines, or specific product-related hobbies. This is definitely the right approach.

You probably should master your metrics...

You mentioned that four ad sets got sales, which is great. But "getting sales" isn't a deep enough analysis. To truly understand perfromance and know how to improve, you need to look at the entire customer journey through your metrics.

Here’s how I diagnose a campaign's problems:

• Low Click-Through Rate (CTR)? High Cost Per Click (CPC)? -> This points to a problem with your ad creative or your audience targeting. Your ad isn't grabbing attention or isn't relevant to the people seeing it. You need to test new images, videos, and headlines. You might also be targeting the wrong people entirely.
• High CTR but few Adds to Cart? -> People are interested enough to click, but they lose that interest on your product page. This suggests a problem with your website. Is it slow to load? Does it look untrustworthy? Are your product photos poor? Is the price too high? Is the product description unconvincing?
• Lots of Adds to Cart but few Purchases? -> You're getting them to the finish line, but they're dropping out. This is often an issue with the checkout process. Are there unexpected shipping costs? Is the checkout process too long or complicated? Are you not offering the right payment options? This is where your BoFu retargeting campaign is so important, to bring these people back.

The ultimate metric for any eCommerce business is Return On Ad Spend (ROAS). This tells you how much revenue you're generating for every pound you spend on ads. A 2x ROAS means you're making £2 for every £1 spent. A 1x ROAS means you're breaking even (before accounting for your product costs). You need to know your numbers—your profit margins, your shipping costs, everything—to know what ROAS you need to be profitable.

Once you know your target ROAS, you can make much smarter decisions about which campaigns and ad sets to scale and which to kill. Use the calculator below to get a feel for how changes in spend and revenue affect your return.

You'll need a proper Testing Framework...

So let's bring this all together. Your'e asking about one ad set in one campaign. But the real path to growth is to build a machine that lets you test audiences and creatives systematically. A single campaign is not a machine; it's a single experiment.

Here's what your account structure should evolve into:

1. Campaign 1: ToFu Prospecting (CBO). This is your main testing ground. Start with a daily budget you're comfortable with. Inside this campaign, create 4-5 ad sets. Each ad set should target a *different*, well-researched audience interest or lookalike. Don't mix them.
2. Campaign 2: MoFu/BoFu Retargeting (CBO). This campaign is for your warm and hot audiences. You can start by combining MoFu and BoFu audiences into one ad set if your traffic is low. For example, one ad set could target all website visitors and social engagers from the last 30 days. Another could target anyone who added to cart in the last 14 days. Your budget here can be smaller, maybe 10-20% of your total ad spend.

Inside each ad set, you should be testing at least 3-5 different ads (creatives). This is hugely important. The ad creative is often the single biggest lever you can pull to improve performance. Test different images, videos, headlines, and primary text. Let the algorithm figure out which ad resonates best with which audience.

This structure allows you to have a clear view of what's working and what isn't at every level: which audiences are driving the best new customers, and which ad creatives are the most effective. This is how you find winners that you can scale.

I've detailed my main recommendations for you to implement in a table below. This is the main advice I have for you:

Moving from a simple setup to a more sophisticated structure like this can feel like a big leap. It requires more strategic thinking upfront, but it's the difference between just 'running ads' and building a predictable customer acquisition system. It takes time to analyse the data, come up with new audiences and creatives to test, and manage the whole process.

This is often where expert help can make a huge difference. An experienced eye can help you set up this structure correctly from the start, interpret the data more effectively, and accelerate your path to profitability. If you'd like to go over your account and strategy in more detail, we offer a free, no-obligation initial consultation where we can give you some more tailored advice.

Hope this helps!

Regards,

Team @ Lukas Holschuh