Our goal was to build a large audience (viewing content, adding to cart and initiating checkouts) to then shift to targeting warm and hot audiences to drive sales.
In doing so, we built data and audiences in our ad account while being highly profitable with our cold audience spending.
The very first thing our Atomic Ads Framework requires is testing the three main variables of every campaign separately: audiences, copy and creativity.
We cannot emphasise enough how important testing correctly is!
During the test phase we use Facebook automated rules to make sure each single piece gets the same amount of impressions. During tests, we target cold audiences only.
Please, stop testing creativity or audiences with broad campaigns! So many people do this and then choose the one Facebook delivers all the traffic to, believing it is the winning combination.
It doesn’t work like this! To get the best results, you have to use single-variable testing (just like your high school science teacher always told you) and leave no space for chance.
The other important thing is to test fast and scale faster:don’t burn too much budget on testing.
From our Atomic testing phase, we got the following results:
Best audiences: 2 Lookalikes 5-10%. One built from IG posts saved and the other from people that researched the website and completely Broad with no interests nor lookalikes (CPM here was very low).
Most marketers out there only build lookalikes with basic events such as purchase or add to carts, but there are tons of other actions users do that can be very lucrative for lookalikes as well.
We set up campaigns matching the best audiences with the best copy and creativity that, in general, were those with an emotional approach to the product opposed to coldly talking about its functions.
With such a strong foundation, the scaling phase was pretty straightforward as each single variable was 100% proven to work thanks to our scrupulous Atomic testing phase.
But you obviously don’t get 27X ROAS without doing something special…
In this case, the trick was constant and automatic optimisation using Facebook automated rules that – unlike humans –work 24/7 If this is set up correctly from the start, scaling up is way easier.
Thanks to our Atomic testing phase and past campaigns the owner ran himself, we were confident we could skip the Atomic consolidating phase where we usually make sure results from our tests actually work and we immediately went on to the Atomic scaling phase.
When scaling, we have different techniques depending on products and the niche. In this case, we worked to minimise the number of campaigns we set up.
Here’s how we structured the ad account:
1ST CAMPAIGN: A CBO campaign with the 5-10% Lookalikes from IG saved postsaudience duplicated 4 times, with each ad set having a different conversion window.
Why dothis? Because you’ll get to different users across your very best audience, which means you’ll get to your very best users, maximising profit.
1st adset: CPA X 1
2nd adset: CPA X 2
3rd adset: CPA X 3
4th adset: CPA X 4
5th adset: CPA X 5
Why dothis? Setting very high caps in a CBO campaign means you are telling Facebook that you are willing to spend a lot of money your conversions and therefore Facebook will let you win all the auctions, which means your ads are delivered to the most profitable users. Even better, since this is a CBO campaign, you not only win all the auctions because of high caps, but you also have optimised CPAs and very low costs of acquisitions. This campaign is generally one of the most profitable.
This is how we implement our Atomic Ads framework for our clients. It’s simple: we make sure we get to the most profitable users at the lowest costs possible to maximise profit. You don’t get these results by chance.