If we could only give one piece of advice on creating profitable Facebook Ads campaigns it would be "Experimentation".
Like all digital marketing channels, Facebook involves testing as many niche audiences with as many ad variations as possible. The faster you can experiment and the more tests you run, the better your performance.
Facebook ads have the most comprehensive target options for any digital media channel, so learning what ads are working, and even which audiences are converting within your target groups is essential.
For this reason, we’ve designed 3 analysis reports that give you the data you need to improve ROI through experimentation. They are free to use, and provide insights in 10 seconds - So no more endless days spent with Excel pivot tables!
What we will cover
- How to quickly cut out ads that are losing you money.
- How to identify new profitable audiences from existing campaigns.
- How to discover the best performing creative using image recognition.
- [Case Study] How we used these reports and techniques for a clients selling books online.
1) Remove ads that are losing money
OpenBook's Analyze Campaigns report helps you quickly cut ads that are wasting you money, and find the ones that are working well. Choose the objective you care about using the "Objectives" dropdown, and add a target CPA (cost per action) for the objective you chose.
The matrix (below) plots the Campaigns, Ad sets or Ads according to the cost per action and the number of actions. So, anything to the top right is great, anything to the bottom left is bad!
2) Find new profitable audiences from your existing campaigns
The Optimize Targeting report means you can run campaigns with wide targeting e.g 21 to 64 age ranges, and we will pull back data on how each age/gender/placement is working within that campaign. It makes it super quick to identify audiences that are worth splitting out into new Ad sets. This speeds up your experimentation, and allows you to learn more about the customers who do convert!
3) Discover the best performing creative using image recognition
Optimize Creative is all about working out creatives work best across your ads. This kind of data can drastically improve results. High engagement rates mean lower CPC's. We merge data across different creatives and their parts and even use image recognition so we can give you data on an image level.
A Case Study
Here's an example of an experiment we ran using these techniques.
We were working with an e-commerce company selling books. We wanted to test if we could drive profitable off-site checkouts from their website using Facebook Interest Targeting (Lookalikes already worked brilliantly!).
- The campaign structure included 5 Ad sets, each targeting a different niche interest-based audience. For example, people who were interested in The Gruffalo, or were Parents.
- We targeted 21 to 65, males and females for now.
- Within each Ad set we created 3 Ad variations with different pictures.
After pushing the campaigns live, we checked performance the following day. A few of the Interest targets didn't generate any conversions. Using the Analyze Campaigns report we spotted these and paused them from within OpenBook.
After running the campaigns for a few days, we looked a little deeper. We were generating conversions, but they weren't profitable at this stage. After removing another Ad set that was drastically under-performing, we fired up the Analyze Targeting report.
Within this is was instantly clear that the CPAs for Males was far, far higher than Females. So, we created a brand new (it's important to duplicate and start again for quality score) duplicate campaign. This time, the 2 remaining Adsets only targeted Females.
Another four days of data, and now the campaign is looking much better. The next stage was to see if any creative variations helped improve ROI. With the Analyze creative report we initially compared all ads for the Checkout objective, then looked at CPAs for the different images. It was clear that images containing pictures of real people playing with the book worked far better than standard book creative.
From the data, we then duplicated the campaign a 3rd time with lots more 'real people' creative. By now the campaign was nicely profitable and we moved on to the next experiment!
It's worth noting that after a while this campaign decayed and CPAs increased, which is common on Facebook. A simple duplication of the campaign and ads within it kicks everything off nicely again. After a few of these, you may need to replace the creative again.