Experimenting with 7 Audiences using Facebook Ads

FB Ads: Experimenting with 7 Audiences

I conducted an experiment with Facebook Ads to determine which audience would generate conversions. Traffic would lead to an e-commerce product page.

The Audience

Using Facebook’s audience insights tool, I researched 30 different audiences within the product specific niche. These audiences consisted of competitor sites, magazines, interests, and people. I narrowed this list to 7 audiences after restricting the age to 25-65+ and ensuring that my audience size would be between 250,000 to 500,000.

The Ads

The ad consisted of a single image with 2 lines of copy. In a previous experiment done with the same product, I found that 2 lines of copy received the most link clicks compared to 1 paragraph and 5 paragraphs of copy. Additionally, in another experiment, I found that a single image outperformed carousel or video formats.

Here’s what the ad looked like:

PROJECT INFO.

Experiment Setup

Campaign Structure

One campaign, 7 ad sets and 1 ad per ad set

Audience

US Only, males and females, 25+, 250,000 – 500,00 audience size

Placement

Facebook newsfeed on desktop and mobile.

Budget

$5 a day for 6 days for each ad set.

$10 a day for 6 days on the ad set that performed the best out of the 7

Bid Strategy

Lowest cost, no cap

Optimization

I optimized for traffic, tracking purchases.

Experiment Results

No conversions were generated from any audience, but audience A1 stood out from the rest due to higher engagement numbers, and CTR and lower CPC.

After scaling up A1, these were the results.

Experiment Conclusions

Larger budget: I needed a larger budget due to the high CPC. I wasn’t able to generate enough clicks and enough data to make a solid conclusion about audience A1.
Product Variation: I received some comments where the person expressed interest in the kit but wished there was an additional product added to the set (i.e. a blanket). Hence, I could experiment with different products within the kit itself.
Experiment with the image: I haven’t tested for different images so changing the image may improve the numbers.
Narrow down age: The majority of clicks for the ad came from an older age group 55+ with a few coming from the 25 to 35 age range. I could redo the experiment, targeting the relevant age range.