Facebook is one of the most popular platforms to purchase targeted advertising. No surprises here: after all, it has a massive audience of more than 2 billion people (if we count Instagram users as well), with the widest available reach in terms of countries, gender, and age. All this makes Facebook if not universal as a traffic source, then at least suitable for the vast majority of advertisers.
However, this tool is not easy to use, let alone intuitive. To use it effectively, you need to account for certain peculiarities of this advertising platform. Naturally, careful preparation is key as well, so you should start getting ready for work as soon as you begin planning your marketing activities. This is why we decided to create a sort of cheat sheet for working with Facebook, in which we will try to list all the advantages, disadvantages, and things to look out for.
Actually, that is the reason why we could not bring ourselves to describing Facebook as a universal advertising tool in the introduction. The thing is that the company's advertising policy imposes a set of restrictions regarding application topics. These restrictions often overlap with the laws of the countries where you plan to launch your campaigns.
Among the areas strictly forbidden for advertising are, for instance, microcredit organizations, tobacco products, multilevel marketing, and adult content. This means that any informational or e-commerce applications in these areas will either not be able to launch their campaign at all or will be banned right after the launch.
Dating apps also face restrictions, though not as severe. These require a whitelisting procedure: before starting any advertising activities, the application's owner must fill out a special application and send their project in for moderation. Only in this case will you able to launch campaigns without the risk of being banned.
A similar procedure is also required for casino, poker, and other gambling apps. These have a separate form that you will need to complete.
Alcohol advertising is fully regulated by the local laws and is thus banned in Russia, for example. Note that this is not limited to apps used to sell alcohol. The restrictions also apply to wine diaries and similar informational projects.
You can see the full list of prohibitions and restrictions via the Facebook Help.
So, let us assume that you have already studied the Facebook advertising rules and have passed the whitelisting procedure if necessary. It is now time to prepare your application from the technical point of view.
1. Firstly, you will need to integrate Facebook SDK into your app's code. You are required to have this SDK if you want to officially run ads on Facebook and have all the tools available for campaign optimization. You can find the integration documentation here. We will just supplement it with a couple of additional comments.
Facebook offers several SDK components for installation:
To launch campaigns, the Ads and App Events components are required in the app (these will be covered below).
All the other components do not affect traffic purchases directly, so they are installed at the advertiser's discretion.
2. Secondly, you need to bind your key app events to the Facebook SDK events. This will allow the advertising network to automatically optimize for the necessary events, e.g., purchases or subscriptions. When doing this, it is important to only use the Facebook events from the following list:
3. Installing a tracker integrated with Facebook: Appsflyer, Adjust, Branch, or Kochava. This step is only needed if you are planning to purchase ads from different sources. This will make it more convenient to view all the stats in a single window.
4. Finally, you will need to configure your Facebook developer account. Using this interface, we inform the advertising network of all the app data, give access to the advertising accounts, and integrate the statistics from Facebook with third-party trackers.
To sign up for a developer account, you will need a social account — a regular personal Facebook page. It is preferable to use a page with some history to it, not a newly registered one. This will help you avoid preemptive banning at the stage of registering your app in the network.
Finally, when all the preparations are complete, we can go on and launch an advertising campaign. To assess its effectiveness correctly, it is important to understand the optimization principles of Facebook, the structure of advertising campaigns within it, as well as some peculiarities of working with this advertising network.
Facebook uses an auction system to determine what ads to show and to what users. When a certain user fits the target audience of several advertising campaigns at once, Facebook compares the ads using several criteria. Based on the assessment, the auction is won by the ad that will have the maximum benefit for the user and, consequently, increase the chances for the advertiser to obtain high-quality traffic.
Facebook lists the following assessment criteria:
● Bid — the sum an advertiser is ready to pay for conversion;
● Estimated action rate — an estimate of the chance that a user will take interest in an ad, based on the data previously collected by the network;
● Ad quality — the quality of a banner, video, or copy and its compliance with the platform's requirements.
All of the above are assessed as a whole. This means that just increasing your bid, for example, will not improve installation conversion if the ad you are using is not good enough by Facebook's standards.
Facebook offers quite a broad scope of opportunities for advertisers to manually configure their campaigns based on gender, age, location, and interests, as well as by uploading a similar, or lookalike, audience (for example, a list of IDs of the users who have already made a purchase in the app).
At the same time, Facebook has developed a system for smart automatic optimization. The network is able to find the optimal bid and target audience on its own, in a process of self-supervised learning. When working on relatively new projects for which there is not yet enough data collected about its paying audience, this is the option we recommend.
Since during this automatic optimization, the system selects the bid and audience by itself, the main bulk of the work done by media buyers focuses on developing approaches (ad concepts) and advertising materials themselves. This is a continuous process since no single approach can have stable effectiveness for a long period of time. So, when a certain approach loses its effectiveness (which takes from a few weeks to a few months), you need to have new ideas and ads ready at your disposal.
You should not be afraid of high bids and low conversion rates in the first days after the launch: the system needs time to learn. The Facebook Help mentions 7 days as the period needed for auto-optimization. However, it has been measured that you can make preliminary assessments on campaign effectiveness as soon as 3–4 days after the launch (but not sooner than that).
When optimizing campaigns for a certain event, the system works a bit differently. It needs to analyze at least 50 events to make conclusions about the appropriate audience and bids. For this reason, we do not recommend choosing delayed events for optimization. An example here would be the situation when you invite your users to subscribe for a one-month trial and charge them only after that trial ends. In this case, it is better to first optimize for the trial subscription and continue working with it until you have 50 purchases.
To assess your campaign effectiveness, you have the following ways to monitor traffic:
● In a third-party tracking system integrated with Facebook;
● In your ad account directly (where you launch campaigns from).
With this, keep in mind that it is normal for third-party trackers to have discrepancies with your ad account of up to 10%. Also, your ad account will have more settings. Actually, the more contractors you have buying ads for the app and the more sources are used, the larger will the discrepancies be, reaching up to 60–70%. Why does it happen?
It mainly has to do with the principles of attribution used in trackers and Facebook ad managers.
Attribution is a process of measuring installations or other events in your app. An attribution window is the period between when a user interacts with an ad (views or clicks it) and the moment they install the ad during which such installation is considered as having resulted from the advertising.
The standard attribution window in the Facebook Ads Manager is 1-day view and 28-day click (usually shortened to "1-28"). This means that Facebook will count a certain installation as a part of this advertising campaign if it happens within one day after viewing the ad and/or within 28 days after the user clicks the ad.
In third-part trackers, attribution windows may be different, which is what causes the discrepancies.
Trackers measure multi-channel attribution based on the latest click. This means that they count a certain installation towards the source that was the last one that the user clicked before the installation. Facebook does not have any idea about what is happening in other sources, so it attributes all the installations that happened after viewing/clicking ads on Facebook to itself.
Example: A user is viewing an ad on Instagram, clicks on it incidentally, and then immediately closes the store. They then go to TikTok and see an ad of this app again. This time, they get interested, click on it, and install the app. In this case, Facebook will count the installation towards itself, while the tracker will attribute it to TikTok because, from the tracker's point of view, the installation resulted from the TikTok ad, which Facebook does not know about.
Facebook measures the date of a certain installation as the day of viewing/clicking the corresponding ad, while trackers measure it as the date the user opens the app for the first time. This can lead to increased discrepancies when viewing statistics for short-term periods.
If you are experiencing significant discrepancies when viewing statistics for a certain date, it is a good idea to first check the time zones used.
When experiencing significant discrepancies in event statistics, check whether the process of sending these data to Facebook SDK and your tracker's SDK is configured correctly.
If the same conversion type is measured by the SDKs of Facebook, your tracker, and a couple of analytical systems simultaneously, this can lead to the Facebook's data doubling, tripling, etc. This happens because when Facebook receives information from external servers, it does not filter it in any way but just adds it to its own data. There are 2 ways to solve this:
Summarizing all of the above, we have put together a checklist. You can use it as a reference when preparing for your work with Facebook to make sure you do not miss anything.
● Study the Facebook help materials to check for restrictions against your app's topics. Undergo whitelisting if necessary.
● Determine your preferred tool for monitoring traffic — a third-party tracker or the statistics of your ad account. If you have chosen to use a tracker, connect it.
● Integrate Facebook SDK into your app, with the Ads and Events components as mandatory, and the rest as optional.
● Bind your key app events to the standard Facebook events.
● Create and configure your developer account.
● When using a third-party tracker, integrate it with Facebook, following the instructions provided in the tracker's documentation.
● Start preliminary assessment of the traffic not sooner than 3–4 days after launching your campaign.
A monthly roundup of mobile marketing news and tips
The importance of A/B testing has been covered in hundreds of texts, books, and webinars. Today, no exponential growth is possible without proper experiments conducted. Let’s figure out whether it’s reasonable to invest resources in split tests if your company is much smaller than Google or eBay.
How to prepare an app and launch an advertising campaign on Facebook? Checklist at the end of the article.