Relevance has always contributed to the cost of reach and desired actions. Now Facebook has taken their internal relevance score and given this metric to advertisers to help them to make more informed decisions regarding optimization. Relevance Scores function similarly to Google’s Quality Score. Below is a recent ad campaign with a new column for Relevance Scores:
A Few Things To Note Regarding Reporting
- Only individual ad units are assigned Relevance Scores. Relevance Scores are not aggregated across the ad set, or the campaign level.
- Relevance Scores are not relevant to organic content. This is a paid metric only. The score is only based on the response your paid reach generates.
- Relevance Scores are calculated after 500 paid impressions.
How to Use Relevance Scores
Relevance Scores are not the end all, be all metric of Facebook advertising. Cost per desired action is still king. Although Relevance Scores are a good predictor of ad success, they should never be your primary KPI. After analyzing countless ad sets, the Relevance Scores do correlate well to the cost per desired action. However, this is not always the case. There have been a few 8/10 scores that are performing at a lower cost than their 9/10 counterparts. Optimize your ads with the same caution as usual. That being said, there are some great advantages of Relevance Scores.
Clients are not just on Facebook for a quick boost in sales. They are there to build brand equity on a platform with high-quality content. Several of our clients promote their everyday content to their fan-base in order to increase the engagement for their Facebook communities. Although we typically judge this evergreen content using Cost-per-Engagement (CPE) and Engagement Rate, Relevance Score serve as another critical metric to find what content is resonating with our community. Since Relevance Scores are based on negative feedback as well as positive, they add a new dimension to the standard metrics that only factor in the quantity of positive engagement.
As I mentioned earlier, this score strongly correlates with the performance (cost per desired action). Therefore, a 9/10 ad unit should yield a high performing ad. Obviously some targets are more expensive than others. And some ad content is inherently more I engaging than others. Which leads me to my next use case: defining and managing expectations.
For example, let's say you need to get under $50 per lead. You have a specific audience that you are going after, and a 9/10 ad is yielding $100 per off-site conversion. It might be time for you to rethink your targeting criteria, creative sequence or test variables downstream of the click (i.e. landing page variables).
Another example is if you have a new product to market, and you want to use a video ad to spread awareness. This product is specifically designed for men in their 30's who earn high income and are single. Your last video ad targeted woman who have similar characteristics as this target. That campaign yielded 100,000 views at $0.05 per view and received a 7/10 Relevance Score so you make the KPI $0.05 per view to this audience. After running the video to the male only target, you receive $0.10 per view but your Relevance Scores is 9/10. This could mean that your video was not as good and Relevance Scores are meaningless, or this could mean that your target market, or even the marketplace at that time was more competitive than the previous targeting/timing. This data will help for managing expectations the next time you go after this target.
How are you using Relevance Scores? Share your thoughts in the comments.