Solved: Is it better to separate Facebook & Instagram ads?

Hi there,

Thanks for reaching out!

That's a really common question, and you're definitely not overthinking it. It's one of those things that seems simple on the surface, but the way you handle it can make or break your campaigns, especially when you're aiming for higher sales. I'm happy to give you some initial thoughts on this. The short answer is yes, you are probably right to want to separate them, but the real reason why your ads might not be performing as well as they could be goes a lot deeper than just a tick-box in the ad set settings.

Let's unpack this a bit, because getting this right is about more than just placements; it's about control, understanding your customer, and forcing the platform to work for you, not the other way around.

You're right to question the "all-in-one" approach...

First off, your gut feeling is correct. Lumping Facebook and Instagram together into one ad set using Meta's 'Advantage+ Placements' is often a lazy way to advertise. Meta pushes it because it's easy and it allows their algorithm free reign to spend your money where it can get the cheapest impressions. But cheap impressions and actual sales are two very, very different things.

The core problem with this approach is that you lose control and, more importantly, you lose data. How can you know if your beautiful, image-led carousel ad is killing it on Instagram but bombing on the Facebook feed if the data is all mashed together? You can't. The algorithm might decide that Facebook is cheaper to show ads on today, so it'll pour 90% of your budget there, even if your ideal customers are all scrolling through Instagram stories and are far more likely to buy there. You end up making decisions in the dark.

By creating separate ad sets for each placement (one for Facebook, one for Instagram), you achieve a few critical things:

-> Clear Performance Data: You can see with 100% clarity which platform is driving results. You'll have a clean view of the metrics for each. Is your Cost Per Purchase on Instagram £15 and on Facebook it's £50? Great, now you know to either kill the Facebook ad set or, better yet, figure out *why* it's failing.

-> Budget Control: You decide how much to spend on each platform. If you know from your data that Instagram delivers twice the Return On Ad Spend (ROAS), you can allocate your budget accordingly, instead of letting Meta make that decision for you based on its own opaque logic.

-> Tailored Creative: This is a big one. People behave differently on Facebook and Instagram. Instagram is highly visual, fast-paced, and story/reel-centric. Facebook is a bit slower, people might read longer captions, and the audience can be older. A great ad is one that feels native to the platform it's on. Trying to make a one-size-fits-all ad for both platforms usually means you end up with something that feels slightly out of place everywhere. Separating them lets you run a stunning video reel on Instagram and a more detailed carousel or single image ad with compelling copy on Facebook. We did this for a women's apparel client and the results were stark; tailored ads for each platform were a big part of how we hit a 691% return for them.

We'll need to look at your campaign objective first... this is where most people get it wrong

Before we go any further down the rabbit hole of placements and ad sets, we need to address the single biggest mistake I see people make. It's choosing the wrong campaign objective. If you tell Meta your goal is 'Reach' or 'Brand Awareness' because you want more people to see your brand, you are actively paying them to find non-customers.

Think about it. You've given the algorithm a command: "Find me the largest number of people for the lowest possible price." It does exactly that. It goes out and finds the users inside your target audience who are least likely to click, least likely to engage, and definitely least likely to ever buy anything. Why? Because those people's attention is cheap. No one else is bidding for them. You're effectively asking the world's most powerful advertising platform to find you the worst possible audience for a sales campaign.

If your goal is "higher sales," your campaign objective MUST be 'Sales' (or 'Conversions' in the old interface). No exceptions. This tells the algorithm to ignore the cheap impressions and instead go and find the people within your audience who have a history of clicking ads and, crucially, making purchases. Yes, the cost per impression will be higher. But you're not selling impressions, you're selling products. This one change is often the most powerful lever you can pull.

I'd say you need a proper testing structure...

So, we've established you should be running a 'Sales' campaign and you should be separating your placements to start with. The next step is to build a proper, methodical testing structure. This is how you move from guessing to knowing. This is how you build a predictable sales machine.

The structure I use for virtually all eCommerce clients is 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. You're targeting cold audiences – people who have never heard of you before. This is where you test your different interests and lookalike audiences.

-> MoFu (Middle of Funnel): These people are aware of you but haven't taken a key action. They might have visited your website, watched a video, or engaged with a post. You're trying to bring them back.

-> BoFu (Bottom of Funnel): These are your hottest prospects. They've added a product to their cart, or initiated checkout, but didn't complete the purchase. Your goal here is to get them over the line.

Here's what a simplified campaign structure might look like to start with. Notice how we are separating not just placements, but also the stages of the funnel.

You'll notice for the retargeting (BoFu) campaigns, I've suggested Advantage+ placements. Why the change? Because these audiences are small and highly specific. You want to be able to reach them wherever they are, as quickly as possible. For cold prospecting, you need the control. For hot retargeting, you need the reach. After a few days of running this, you'll have clear data. You'll see which interests work, and on which platform. Then you can start turning off the losers and scaling the winners.

You probably should focus on who you're targeting...

A perfect structure and placement strategy is useless if you're showing your ads to the wrong people. The most common mistake I see here is targeting interests that are way too broad. If you sell high-end coffee beans, targeting an interest like "Coffee" is a waste of money. You'll be reaching millions of people who are perfectly happy with their instant coffee. Your ad will be irrelevant to 99% of them.

You need to think about your customer's pain, their nightmare. What problem are they trying to solve? For the coffee lover, it might be the "nightmare" of a bland, boring morning cup. Your targeting should reflect that. Instead of "Coffee", you should be testing interests like:

-> Specific high-end coffee machine brands (La Marzocco, Sage, Rocket Espresso)
-> Famous specialty coffee roasters (your direct competitors)
-> Coffee publications or influencers (James Hoffmann, Perfect Daily Grind)
-> Behaviours like 'Engaged Shoppers' layered with those interests.

This is about finding proxies for intent and quality. You want to find audiences where the highest possible percentage of people in it are your ideal customer. Once you have some sales data, you can then build Lookalike audiences, which are often the holy grail of prospecting. A Lookalike of your past purchasers is an incredibly powerful tool. Here's the general priority I follow for audiences:

1. BoFu/MoFu Retargeting: (Highest intent) Website Visitors, Add to Carts, Initiated Checkouts. This is your lowest hanging fruit.
2. Lookalike Audiences: (Scaling winners) Start with Lookalikes of your best customers, then purchasers, then work your way down the funnel (Add to Cart lookalikes etc).
3. Detailed Targeting (Interests): (Finding new customers) This is where you do the hard work of testing and finding new pockets of customers, using the specific, problem-aware approach I mentioned.

And what will it all cost?

This is the million-dollar question, isn't it? The cost of getting a sale varies wildly based on your industry, your price point, and the country you're targeting. But to give you a rough idea for eCommerce in developed countries like the UK, here are some ballpark figures we often see before we start optimising.

So, you could be paying anywhere from £10 to £75 for a single sale. If your product costs £40, a £75 cost per acquisition is a disaster. If your product costs £500, it's fantastic. The entire purpose of the methodical structure I've outlined—separating placements, optimising for sales, smart targeting, funnel-based campaigns—is to push your costs towards the lower end of that spectrum and maximise your Return On Ad Spend (ROAS). This is how we've managed to get results like a 1000% ROAS for a subscription box client, or an 8x return for a brand selling maps. It's not magic, it's just a rigorous process.

I've put my main recommendations for you into a table below to make it as clear as possible.

This is the main advice I have for you:

As you can see, the initial question about separating platforms opens up a much bigger conversation about how to advertise effectively. Getting this stuff right is a full-time job, and implementing a structure like this, creating the different ad versions, and monitoring it all daily is time-consuming and requires a fair bit of experience.

This is the kind of methodical process we implement for all our clients. It's about moving away from guesswork and building a system that predictably generates sales. If you're serious about growing your sales and feel like you could use a second pair of expert eyes on your campaigns, we offer a free, completely no-obligation strategy session. We can have a look at your ad account together and I can give you some concrete, actionable advice based on what we see.

Hope this helps!

Regards,

Team @ Lukas Holschuh