It’s important to understand the purpose and process behind how Facebook “learns” to target your ads. Here’s an overview.
When you create a new Facebook ad, there will be an initial “learning phase”. Here’s what you see:
So what’s going here, and how does it affect your approach to creating ads?
Algorithms At Work
The Facebook advertising system is, in fact, a computer algorithm.
At a basic level, an algorithm is just a series of steps for solving a problem. But of course, systems like Google and Facebook use massive computational power that solves problems with millions of variables.
On Facebook, we want the algorithm to put our ads in front of just the right people at the best possible moment. Its purpose is to hone-in on the people most likely to take action on your ads.
The Learning Phase
So how does Facebook “learn”? What is it learning?
To display an ad as effectively as possible, Facebook correlates two main pieces of information: the goal of your campaign and target audiences. Its goal is to maximize your ad’s conversion and optimization events.
When the campaign starts, Facebook doesn’t have the behavioral data to deliver your ads with maximum accuracy. So it runs through this initial learning phase, showing your ad to different types of people and getting data on the first 50 optimization events that occur.
When it has this data, it gains a higher level of understanding about what type of person within your audience sets is most likely to complete your goal.
After the learning phase is complete, the power of the algorithm really shines. It can now pick, on a person by person basis, the people who are most likely to convert. It’s not just targeting a demographic or broad interest. It’s targeting individuals based on specific, personalized data sets.
Know Your Goals and Your Audiences
The Facebook algorithm is delivering ads with computer power no human brain could match. It’s a powerful tool.
But it’s also working off important criteria that you set, which is your campaign goal and target audiences.
Setting the right campaign goal is a vital step because Facebook will explicitly target people based on that goal. Here’s an overview of the goals you can target:
Say you’re trying to develop brand awareness with a video, so you target video views. When Facebook is going through the learning phase, it’s discovering people in your ad set targets who tend to watch videos on Facebook. It’s targeting that behavior as precisely as possible.
If, however, your business goal is demand generation and you’re targeting website conversions, it will target people most likely to engage in that conversion event.
In both of these cases, you could have the same target audiences. But because of the goal you designate, Facebook will target different people, seeking the individuals most likely to give you favorable results.
When you start a campaign, you must know what you want to achieve and who you’re trying to reach. Your campaigns are driven by your business objectives, and ad content is driven by your target audiences.
With budgets set at the campaign level, Facebook will automatically split test your ad groups to deliver to the best audiences. After it goes through the initial learning phase, you can turn off audiences that aren’t performing and move your budget to where you’re getting results.
Let the learning phase run its course when you start a campaign. If you make edits, it starts the cycle over and you lose the data it’s gathered to that point.
Also, be careful to not set budgets too high. With a lower budget, you force Facebook to find the best audiences for your ad sets.
When all the learning is done, you can put your budget and content towards achieving your goals with maximum efficiency.