How I increased ROI with 1789% using Facebook & Instagram ads

In a 2-month period, I increased the monthly ROI with 1789% of this e-commerce stores Facebook/Instagram campaigns. The key to this strong growth was extremely active management of Facebook Ads with multiple A/B testing.
The company, NicheForBeach, sells premium beach-wear accessories on the US market. All their products are eco-friendly and handmade.
Sounds pretty good right?
It’s really simple. Let me explain how I was doing it…
There are two steps I am using that applies to create profitable campaigns like this (that apply to virtually any industry):
1. Broad testing of ads with a budget at ad sets level.
2. Optimize best performing ad sets with CBO campaigns (campaign budget optimization).
Want to apply these strategies to your own campaign?
In this article, I’ll walk you through the campaign, what tools I’m using, and how I set them up so that you can create your own success story.

Campaign Overview

Here’s a quick walkthrough of the Campaign strategy:
Step 1: Single interests Ad Set Testing
Step 2: Combine interests Ad Set testing + Ad Testing
Step 3: First CBO Campaign
Step 4: Second CBO Campaign

Testing for all you’re worth!I like to test a lot. And I mean a lot! This gives me better data to make the right decisions about which interest to use.
When testing interest, I recommend breaking it down into five categories:
10 interest with Brands (ex Victoria Secret)
10 interest with Communities (ex Sustainable Products, Bohemian Style)
10 interest with Behavior (Ex Travellers)
10 interest with Products (Ex Sandals)
10 interest with Uncategorized (Ex Summer, Sea)
All ad sets only has 1 interest each. This is NOT the step where you want to combine interest. The goal of this step is to collect data. You pay Facebook for data, not results.
Depending on your budget you can adjust the number of active ad sets. However, I recommend using a minimum of $2 per ad, so if you have on ad set with 4 different ad inside it, use a minimum daily budget of $8.
After testing different audiences, the next step is to combine the best performing ones.

Combining Lookalike and interests-based targeting:One key strategy I am using is combining Lookalike audiences with interest-based targeting.
Why?
To filter out 90% of the audience, and only work with the users who are most likely to make a purchase.
First, we select a Lookalike audience from the website visitors of 10 % (with the data collected from previous steps. That’s why testing more interests will broaden your audience and give you better results) and we now have a potential reach of 21 million.
Good, this way we have already sorted out 90% of the audience.
With a new pixel, without massive amounts of data 10% lookalike is a fair population to work with. Not too broad, and not too small to risk excluding potential buyers.
Since we were selling sandals, the next step is to add sandals as an interest to this Lookalike audience. This cuts down the audience to 6,4 million people. Notice something? Look at the graph below to understand the mechanism behind this.

It’s also possible to narrow down the audience with additional interest that you already have tested to be super specific in your targeting.

With this principle in mind, we can keep narrowing down the audience even further.
This is why step 1 is so important!
Instead of having 5 interest in one ad set and you guess which of the ad sets that performs the best, you are now 100% sure which interests to use for best performance. Single interest ad sets are the way to go when collecting data.
Make combinations with and without Lookalikes as the Audience base. Then select the top 5 audiences and ads and insert them into one CBO (Campaign Budget Optimization) with 5 ad sets.
When you have clear data of the winner in this Campaign, you duplicate this Ad set into a one CBO Campaign, but with 5 identical ad sets. Optimize placements, age, devices, gender, states, region, etc… And Let Facebook’s algorithm do the work.
This is the best way to make broad testing and narrow the audience step by step. You will not miss any valuable customers and this is the most effective way to quickly find the audiences you should be targeting.

With this principle in mind, we can keep narrowing down the audience even further.
This is why step 1 is so important!
Instead of having 5 interest in one ad set and you guess which of the ad sets that performs the best, you are now 100% sure which interests to use for best performance. Single interest ad sets are the way to go when collecting data.
Make combinations with and without Lookalikes as the Audience base. Then select the top 5 audiences and ads and insert them into one CBO (Campaign Budget Optimization) with 5 ad sets.
When you have clear data of the winner in this Campaign, you duplicate this Ad set into a one CBO Campaign, but with 5 identical ad sets. Optimize placements, age, devices, gender, states, region, etc… And Let Facebook’s algorithm do the work.
This is the best way to make broad testing and narrow the audience step by step. You will not miss any valuable customers and this is the most effective way to quickly find the audiences you should be targeting.

First CBO Campaign

Facebooks Campaign Budget Optimization function has become extremely powerful in the last year. The next step is to insert the top 5 performing ad sets into 1 CBO campaign. What you are doing now is letting Facebook’s algorithm and machine learning do its job.

Wow… this machine knows how to do a good job!

The only thing you are doing in this step is to keep optimize the campaign. Even if Facebook will do most of the job automatically, why not give her an extra push? Narrow down placements, devices, age, gender, regions, etc.

Whenever you can clearly see a pattern of which ad set performs the best in this “elite-Campaign” you duplicate the ad set it 5 times into a new CBO-Campaign.

Second CBO Campaign

If you ever felt like it’s the same effect when increasing Facebooks budget as putting gasoline on a fire, don’t worry!

When you increase the budget on CBO campaigns the learning phase will not reset. This will make the scaling process a comfy ride.

You will have 5 identical ad sets, and you are giving Facebook all the control of optimization. The mechanism behind the reason for activating 5 identical ad sets is to not risk a “bad start” since this will affect the whole campaign. Now you are statistically minimizing the risks and can ride the wave with the best performing ad set.

Look at the example below after the first 14 days of doing this to one of the products.

The ResultsThis case study is written after 2 months of managing the Facebook ads of this e-commerce store.
The ROI increased with 1789% month 2.
The total increase of ROI (2 months) including the testing month is 330%.
As you can see, you pay for data. But you also pay for long term success since the return will be massive when using this strategy.
Facebook is a beautiful money-making machine if you treat her right.
If you do one step before the other, she might give you a hard time.
Treat her right, and she will give you what you want.

Do you want results like this?

Connect With Us Today