Neither the strongest nor the smartest survives in life. The one who can adapt to change the most is the one who survives.
Charlis Darwin
The Meta company’s applications have how many monthly users?
- Facebook: 3.03 billion
- Instagram: 2.35 billion
- Messenger: 1.3 billion
- WhatsApp: 2.7 billion
The data are monthly active user numbers for the year 2023.
So, 9.38 billion data flows live to Meta every month. Meta dominates everything from demographic information of these users to what they like, when they are happy, and when they are unhappy, even to their daily behaviors!
When it comes to Meta, they usually call it “The All-Seeing Eye.” Because it knows about every step we take.
Meta needs to control this data, and it does so with its own technologies and algorithms it writes. Of course, for simplicity, let’s just call it the algorithm for now.
Why does it do these? Because in 2023, Meta made $134.72 billion in advertising revenue. This figure was $121.9 billion in 2022.
Looking at this reason, it will be very understandable that organic engagement, which was 5.5% in 2019, is now around 0.25%.
As a result, Meta needs advertisers. It is estimated that it earns 99% of its revenue from advertisers. If advertisers cannot achieve the performance they want, they will not give. Facebook knows this best because this situation means a decrease in revenue.
So the most important job of this algorithm is to show the ad to the right person. It has been developed for this purpose since its establishment. Let’s take a look at how it has evolved without boring you.
Evolution of Facebook
2005: Facebook was founded.
2006: Newsfeed was created.
2007: The “Like” button was introduced.
2009: “Sorting order” Posts with the most likes began to rise to the top of the ranking.
2015: It began to lower the ranking of pages with high volume promotions. t introduced the “See First” feature, which allows a page’s post to appear instantly in your feed when it sends a post.
2016: It added the “Time Spend” ranking system. It began to measure based on the time a person spends with it, even if users do not interact with the post. It introduced the feature to start live broadcasts.
2017: “Emotional reactions” More than just Like.
For videos: Introduced the “Completion Rate” ratio. Videos that people watch until the end started to be shown to more people.
2018: Zuckerberg: “posts that spark conversations and meaningful interactions.”
He announced that he would prioritize posts that encourage conversation and interaction.
2019: It started to show the posts of your tagged friends you talk to in DMs as “Close Friends” by defining their posts to you more prominently.
2020: It announced that helping users understand the algorithm and control their own data would help provide better feedback to the algorithm. Concerns started to rise. Additionally, it announced that it would evaluate the reliability and quality of news.
2021: “Score system” Based on likes, saves, and comments, a system designed to discover new content.
You can check it out here: https://www.instagram.com/p/CRZsz1whj5e/
Now let’s move on to how the algorithm evolved in 2024.
The main thing you need to know is that Artificial Intelligence (AI) has become an integral part of the Facebook algorithm. With its advanced machine learning capabilities, AI helps personalize content for users by analyzing their behavior and preferences on the platform.
As of 2024, the Facebook algorithm determines which content is most relevant to a particular user based on four main ranking signals:
- Who posted it: You’re more likely to see content from sources you interact with, including friends and businesses.
- Type of content: If you most often interact with video, you’ll see more video. If you engage with photos, you’ll see more photos.
- How likely you are to engage with the post: If the algorithm thinks you’re likely to comment, like, or share a post, then it’s more likely to show it in your feed.
- How interested will other people be in this post: If the Facebook algorithm determines that a post will generate meaningful interactions between users, such as comments, shares, and replies, it will be ranked higher.
Source: https://www.facebook.com/business/help/718033381901819?id=208060977200861
Now that we understand how the algorithm works, how should we, as advertisers, look at the subject?
This algorithm is actually a “natural selection engine”.
“Natural selection” is actually the main process of evolution, there is a lot of information on the internet. Natural selection exists in all organisms, in all living things; it is like the main theme of evolution; those that best adapt to their environment survive! Those who cannot adapt die like dinosaurs.
We can think of the Facebook algorithm as similar to “Natural Selection”; the algorithm kills everything that cannot adapt to its environment without batting an eye.
We have an entrance and an exit. We have made the advertising entries into the algorithm, we can call it an environment full of people, like a pool, the better we adapt to this environment, the better our output will be, which reflects on our performance.
The Meta algorithm includes all advertisers’ campaigns; they all compete to survive.
I think many amateur advertisers do not understand how this algorithm works. What we want does not matter at all. “Meta” gets what it wants. Our task is to give it what it wants.
The 5 main points that the algorithm expects from you:
- Diversity
- Customer journey
- Facebook ad policies
- Selecting the right campaign objective
- Performance Management
Diversity in Ads
In messages, images, and target audiences.
The algorithm combines and tests millions of times every day the data that all the scientists on Earth would analyze in a year through these 3 variables. This machine is already designed to test and analyze variables.
If we try to run a campaign with just 1 image, 1 text, and a small group of target audience, it will most likely be a failed campaign, and we will not spend the customer’s money correctly. As a result, we will not achieve the performance we want.
Also, this is not strategically sound; the algorithm gives a score to the ads that are weak as soon as they come out of the learning process, it codes them as unsuccessful from the beginning, and how can you win with just one image in the same race as millions of advertisers?
Customer Journey
You have diversity in texts, images, and target audiences, but you couldn’t position it like a journey. In other words, you couldn’t get them into a story. Still, this diversity won’t attract them. One of the misunderstood issues is this, what is the attractiveness that will take action for the target audience?
Maybe the Instagram account is very boring. Or the website/landing page is not persuasive enough. Maybe the CTA button on the website is not used in the right places.
Remember, the algorithm can determine what the target audience likes and dislikes! Let’s say, you took action on 70% of the audience who saw your ad, brought them to the website, but they took no action on the website, left after 5 seconds. Boom! Cost per click goes up. The result is bad end-of-month reports.
Facebook (Meta) Advertising Policies
If we do not comply with advertising policies, it doesn’t matter how good our journey, text, image, or diversity of target audience is. Because when you don’t comply with the policy, the ad display is stopped immediately, and if you repeat it, your ad account is banned. Then struggle.
Campaign Objective
Campaign objective affects the entire performance of the ad, but it should be decided at the beginning according to the target the brand wants to achieve at the end of the campaign. Each objective affects the way the algorithm works. I will explain these in detail in another article.
Performance Management
Finally, the algorithm expects you to manage your campaigns correctly. After the learning phase ends, the Meta algorithm determines the best-performing text, image, and target audience combination. This is a process of evolution, and the algorithm expects you to be involved in this process and manage it correctly.
One reason why some ads fail is that they either close the ads early or may have detected the problem too late. One of our jobs is to give the algorithm enough budget to do its job well. Convincing the customer is still part of our job.
Thanks to those who have read this far! It may be boring because it is one of my first articles. I will improve myself. Goodbye for now.