Lookalike audiences in Facebook marketing - Blog Dont.Farm
Imagine an appliances shop. Somebody enters in a need of a vacuum cleaner. He or she has not decided on the model yet, doesn’t know which type to choose, but has already read a lot of articles on the Internet and is quite certain that the purchase needs to be done today. A shop assistant greets him or her, and in 15 minutes, the purchase is complete. Huh, that’s strange what just happened. After all, this person was planning to visit three more shops before making a decision. The secret is that the assistant knew that they wanted to make their purchase today!
It would be nice if all sales proceeded according to the same scenario. But how do you achieve that? How do you attract customers? By showing them targeted ads! Well, how do you do that? In the offline business world, this is impossible because ads here are seen by everyone, both those who decided to buy a vacuum cleaner and those who could not care less about hoovers. And while segmentation using niche magazine works (sort of), in case of banners and billboards, most of those who end up seeing such ads are what marketers call a non-target audience.
However, online marketing is very different. You can show ads to precisely the users who have already visited your site and even performed a specific action. Such tools are offered by a handful of services, and Facebook is one of them.
Let’s start off with the definition. As the word “lookalike” suggests, this approach involves some kind of comparison. But what kind?
Let’s assume you have a database of e-mail addresses of users who have performed some particular target action on your site. For example, they bought something there.
What you do now is you upload this into Facebook and allow their algorithms to match your database with the database of all Facebook users to identify similar characteristics.
To give a somewhat far-fetched example, imagine that you are selling yoga mats. Facebook neural networks find out that most of your users are members of groups dedicated to healthy eating and DIY home improvement projects and are not interested in politics. The analysis is done, and you get your target audience with similar characteristics.
Lookalike audiences are a powerful tool which does not have any “manual” alternatives. Of course, hypothetically, you could track your users’ behaviour yourself. However, keep in mind that all you have are their e-mail addresses. What are you going to do with those? You cannot run any kind of analysis without special software. And even if you buy it, it would not be very effective compared to neural networks. These algorithms are configured to constantly improve, and marketers claim that they do become more effective month after month.
You don’t need to know how this lookalike technology works. And you can use it for your purposes absolutely free:
Go to the Audiences section, which can be easily found in your Facebook Ads Manager;
Click the Create Audience link;
Choose the Custom Audience if you want to import the list of users that have already been analyzed; you can also choose Lookalike Audience if you have already put together your own database.
Done! You now only need to choose the degree of similarity. The audience should be as large as possible. You cannot achieve anything with just a couple hundred users.
The resulting database that Facebook will match for you is at least 1000 people. Notice that these are not surefire buyers or visitors that will go to your site or download your application no matter what. These are just the users your ads will be shown to!
So the database should be as big as possible – ideally, thousands and even tens of thousands of people. This will help the platform to find your genuine target audience, which you will see for yourself by monitoring the increase in your conversion!