Cubatica is a marketing agency in Australia that serves various clients with budgets ranging from $100,000 to over $1 million in ad spend per month.
One of the company’s departments works with an in-house team of traffic managers who run offers as affiliates, including being top affiliates on various networks.
The company already runs several successful campaigns, routinely generating over $100,000 in profit every month.
The challenges:
- The company already generates significant revenue with an ROI of 30% to 50%, so the challenge is to continue growing.
- Being dependent on the Facebook platform, almost every month there was some banning process, and many accounts were frequently suspended.
When we started this work, the company was already investing an average of $10,000 to $50,000 per day in various offers, reaching peaks of over $100,000 invested per day.
In situations like this, it is not always easy or logical to establish a path to continue growing steadily, especially because there is already an established process that generates consistent results.
Business Manager Contingency Structure
To mitigate the risk of ad accounts being suspended, we need to prepare a special contingency structure.
Instead of having all assets (pages, domains, pixels, etc.) concentrated in just one Business Manager (BM),
we adopted a separation of environments:
- One BM for pixels and domains;
- One BM for pages, with pages being directly shared with the profiles that will run the ads;
- One BM for running the ads;
Additionally, ad accounts are provisioned through external agencies, which increases the protection of the BMs.
This way, we created a structure that doesn’t completely solve the issue of Facebook bans, but it does mitigate a significant portion of the risks.
By adopting this structure, the number of ad restrictions on the pages has drastically decreased (from several blocks per week to 1 or 2 blocks per month).
Of course, depending on the campaign being executed (especially those related to the health niche), ad account suspensions still occasionally occur, but this has also decreased to a monthly interval (whereas before, account blocks were common every week).
To complete this structure, we also replicated the same environment configurations using distinct agencies.
In other words, even if there’s an issue with one ad account, BM, or even an agency, we have the same campaigns running in various setups with the same contingency structure.
Optimizing Tests
The normal advertising process established in the company is to test many creatives in one week to determine whether a specific campaign is worth continuing or not.
Typically, the process followed the following rule:
- Create a landing page;
- Create a variation of ad copy;
- Create 20 creatives and check which ones are the winners;
If the campaign performs well, ads are scaled.
This process has always worked very well, but we began to notice several difficulties.
Facebook operates on an auction process. Depending on the wording of the ad copy, Cost per Click (CPC) and Cost Per Thousand Impressions (CPM) metrics can become very high, and we need to test different variations of copy (and also landing pages).
So we established the following process:
Instead of testing 20 creatives with one landing page and one copy, we adopted the criterion of 3 to 5 creatives, 3 to 5 copies, and 3 to 5 landing pages.
Considering only the minimum number of tests, we already have 27 (3 x 3 x 3) variations of creatives that allow us to have a clearer view if we have a strong campaign on our hands or not.
We often also test more than one offer in parallel, and we can quickly cut the worst-performing ads and compare the final conversion.
Initially, campaigns were tested in countries considered “Tier-1”, which are of higher quality.
Over time, we decreased to 4, and in the initial creative tests, we tested in just one country first, and then tested for others.
The validation time for offers has reduced from 1 to 2 weeks to 3 to 5 days.
Before, it was necessary to invest an average of a thousand dollars, and now with about 500 dollars, we already know if we have a winning offer or not.
Optimizing Scale
After finding a creative with good performance, the scaling process can happen in various ways and configurations, and depending on the time of year, one configuration may work better than another.
After finding the top-performing creatives and being able to maintain profit at scale (creative tests typically run at $500/day, and scale tests start at $2,000/day), we move on to the process of expanding to other countries.
Previously, the company ran a single campaign for all countries, but Facebook always prioritized only one or two, and we couldn’t achieve the scale we needed consistently.
So we started creating separate campaigns for each country and analyzed the times of best performance to determine if the campaigns are worth running 24 hours a day or if it’s better to run them only at certain times of the day.
Depending on the campaign, we also conduct tests targeting different ages and genders, with some pleasant surprises.
This was a lengthy process, but it brought two significant benefits: stability and consistency in scaling.
With Facebook’s numerous advertising restrictions, we need to be creative to continue reaching our target audience.
But remember that Facebook operates on “engagement”.
Even if your audience is more restricted, a message that resonates with them will help Facebook “learn” who your target audience is and show the ads to them.