Facebook Advertising has become the most efficient way to advertise online. Every business owners are taking advantage of Facebook ads. When there is a need to make the campaign profitable, we also need to know how FB ads algorithm work? Only then we can run the campaign profitable.
In this post, I am going to explain How FB Ads Algorithm Work?
Many FB marketers would have heard that ” duplicate the ads 3 times“, “run 1 ad/adset” etc.
Yes, We will see why they are Important?
There are so many aspects of FB Algorithm. Upon them below 4 are the important ones.
- Auction System
- Campaign Structure
- Machine Learning
- Quality Score
Let’s see in depth about these 4 main aspects of FB Algorithm.
No coupons found.
I. Auction System:
First, You need to know there are thousands of advertisers like you. So when you are creating a campaign you are indirectly competing with them but Facebook will try to keep everyone in rejoice.
Sometimes, Facebook will fail to satisfy every advertiser in the competition. So FB Algorithm gives maximum attention to highest bidders. It doesn’t mean your bid should be always high.
Rather than the maximum bid, there are some factors for the better performance of your campaign. They are
- Relevance Score
- Link CTR
- Engagement etc
Based on the above factors facebook algorithm will give better results for a campaign even among a huge number of competitive bidders.
There are many possibilities that a campaign with good relevance score and engagement rate can beat a campaign which is having a higher bid.
Did you ever notice, Why the campaign with the same targeting give good results in some adsets and doesn’t work in another adset?
I will explain why there is a difference. It is because of audience pools.
Facebook separates the audience into different pools based on their purchase behaviors, engagement rate etc. So When you launch a new adset the total audience in the particular targeting will be created as portions or pools.
Let’s see an example:
I had attached the above image just to understand the concept.
In this adset, I have 4 ads. They are 3 duplicates of a 1 ad. They all have the same targeting and they are on same budget. When you start running campaigns you get different result for different ads.
The 1st ad may be performing well with good shares, the second may have average reach but got 2 ATC while the third ad got no engagement likewise. This is because of the audience pools. Each pool may perform differently. You need to run your ad until you find the highest performing pools.
II. Campaign Structure:
This is a very important part. You need to structure your campaign properly to pick the right ad. Based on the ad performance you can analyze and get accurate data. Before diving into this get to know what is campaign, adset and ad.
I will give you a method to structure your campaigns.
- 1 Campaign – x number of adsets with 3-4 ads/adset. ( Highly Recommended )
- 1 Campaign – x number of adsets with 1 ad/adset.
Note: Here x represents any number of adsets. But for better results, you can run 1 – 3 adsets per campaign.
By this way, you can split test and get accurate results for further breaking down the campaigns.
III. Machine Learning:
As you know facebook algo completely works on Machine Learning, the results are driven by relevance score & engagement rate. Facebook gives more attention to your top engaging ads.
Did you ever notice, Why most of the users are from the same country in your winning worldwide targeting adset?
Because facebook algorithm learnt that x country audiences are performing well on your particular adset and they keep sending traffic only to a particular country.
Likewise, you should know about the two main phases of FB’s Machine Learning:
- Learning Phase
- Optimizing Phase
In your learning phase if your conversion window is 7 day ( mostly everyone chooses 7 day becuase of long term run) you need to have at least 50 numbers of your obejective.
Initially, 20 Conversions was said to be enough to run WC Purchase, but now facebook suggested it to be 50. These phases are defined by adset level.
I mean if your objective is VC then you get 50 VC, if your objective is ATC then you should get 50 ATC to complete the learning curve suggested by facebook algorithm.
Tip: If you think you cannot reach your objective with your current daily budget, Increase your daily budget.
Note: When your adset is in learning mode don’t make frequent edits other than your budget.
When your adset finishes learning phase it enters into optimising phase. During this phase, you will get better results on your adsets. If you are not getting better results then your adset needs more attention based on different factors.
Now, the good performing adset in the second phase needs to fed with more budget. This part is also called scaling part. Scaling of Facebook Ads is an art. We will cover this topic in our later article.
Note: You should scale your adsets budget by 25% daily for healthy performance.(Suggested by pro marketers)
No coupons found.
IV. Quality Score:
As you see in AdWords about the quality score of the ad account, Facebook system also have their account quality score but they are not measured like AdWords in metrics. They are conceptual.
If your ad account quality score is good then you will be getting good engagement rates, winning more auctions, getting campaigns approved faster, better traffic etc. FB ads algorithm automatically gives good results for the account with good quality scores.
To Improve your Ad Account Quality Score follow these below steps:
- Launch ads daily, even with smaller budgets.
- Make sure you have good relevance score on your ads.
- Give a prompt reply to your facebook page audience making queries or asking help in comments.
- Always use high-quality domain extensions. ( like .com, .in etc )
- Don’t get more disapprovals, more policy violations.
- Engage more with Facebook Help Center.
- Have prompt billing. Add few extra CC’s to avoid billing issue.
- Embed your Facebook page on your website.
- Follow these in all your ad accounts.
This is how the FB ads Algorithm works. Bring a Share if you like this article. Hope it had come out well.