Who decides whether a person should see an advertisement or another is not a group of specialists dedicated to the subject, it is an Artificial Intelligence system trained to obtain the best results.
Now Facebook published information about how they decide which ads to show to people, highlighting two main points: the target audience selected by advertisers and the results of their ad auction.
The Basics of Facebook Ads
It indicates that everything starts when advertisers choose their target audience through their tools. Audiences are created based on categories such as age and gender, as well as the actions that people take in applications (Instagram or Facebook). An action can be a like to a Facebook page or a click on an existing ad. Advertisers can also use the information they have about their audience, such as an email list or people who have visited their website, to create a custom audience or similar audience.
On the subject of the auction, things change a bit. Facebook selects the top ads to show to a person based on the ads that have the highest total value score – a combination of advertiser value and ad quality. The ads with the highest bid do not always win the auction. Ads with lower bids often win if your system predicts that someone is more likely to respond to them or finds that they are of higher quality. This allows companies of all sizes to compete in the auction and reach customers on any budget.
For this, decisions are made based on:
– Advertiser bid: how much money an advertiser spends on their Facebook ads – Estimated action rate: Facebook estimates the probability that each user will take action on an ad, based on a number of factors related to their individual behaviors, such as and as discussed earlier. – Ad quality: Facebook measures this based on user feedback (for example, how many people report or hide your ad) and low quality attribute evaluations (too much text in the ad image, tabloid language, and other points not recommended when creating the ad).
Machine learning in Facebook ads
Machine learning is a system that learns as it receives new data, without being explicitly programmed. On Facebook it is used to generate the estimated action rate and the quality score of the ad used.
On this matter, Facebook comments:
To find the estimated action rate, machine learning models predict the probability that a particular person will take the action desired by the advertiser, based on the business objective that the advertiser selects for their ad, such as increasing visits to their site. web or boost purchases. To do this, the models consider that person’s behavior on and off Facebookas well as other factors, such as ad content, time of day, and interactions between people and ads.
Examples of non-Facebook behaviors that models consider include things like visiting a website, buying a product, or installing an app. That is, if we install an app outside of Facebook, it will be taken into account when viewing ads inside.
Advertiser value is calculated by multiplying an ad’s bid by the estimated share rate. This is an estimate of the probability that that particular person will take the action desired by the advertiser, such as visiting the advertiser’s website or installing their app. They then add the Ad Quality Score, which is a determination of the overall quality of an ad. From there comes automatic learning.
You can check more details in this Facebook article.