- Basic principles of SKAN3 📚 First, let’s review the basic principles of SKAN. SKAN (StoreKit Ad Network) is an advertising attribution solution provided by Apple, which is mainly used to track whether users install apps after clicking or displaying ads. The whole process can be divided into the following steps:
- Ad display or click: The user sees the ad or clicks on the ad. 2. Installation: The user downloads and installs the app. 3. Apple attribution: Apple uses the SKAN framework to attribute whether the installation comes from ad display or click. 4. Send back to advertisers: Apple sends the attribution results to advertisers, and advertisers optimize advertising based on this result.
Sounds simple, right? 😊But in reality, there are many details and mechanisms that will affect the final allocation result.
- Timer Mechanism: How does the timer affect data reporting? ⏰In the allocation process of SKAN, there is a very important mechanism called the Timer mechanism. This mechanism is mainly used to control the data reporting time.
What is the Timer mechanism? Simply put, the Timer mechanism is a timer. When you click on an ad, Apple will start the timer, but it will not tell you the result immediately, but will wait for a period of time (usually 24 hours) before reporting the data to the advertiser. Apple does this to protect user privacy. If the data is reported too quickly, it may expose the user’s behavior habits, such as when you clicked on the ad and when you downloaded the App. Therefore, Apple chooses to “delay” to make the time of data reporting less accurate. Impact of the Timer Mechanism: Data Delay Because of the existence of the Timer mechanism, the data seen by advertisers is often later than the actual time. For example: Monday: You clicked on the ad and downloaded the App. Wednesday: Apple attributed this download to the channel. Then the data that advertisers see on Wednesday will include the installation volume on Monday. This will cause the data to not match.
- The “window period” of attribution: Who is the attribution attributed to? ⏳Next, let’s talk about the “window period” of attribution. The so-called window period refers to the period from when a user clicks or displays an ad to when the app is installed. During this period, Apple will make attribution judgments. 1. The length of the window period The length of the window period is usually click (30 days) display (1 day). That is, if the user installs the app within 30 days of clicking or within 1 day after displaying the ad, then this installation will be attributed to this ad display or click. 2. The object of attribution The object of attribution is ad display or click. That is, users will be attributed not only for clicking on the ad, but also for simply displaying the ad. This depends on the specific settings of the ad and Apple’s attribution rules.
For example: Suppose a user displays an ad on January 1, but installs the app on January 30. Then this installation will be attributed to the ad display on January 1. The data that advertisers see on January 30 will include the installation volume brought by the ad display on January 1. This will cause the data to not match. - Privacy Threshold: How does it affect data reporting? 🔐In addition to the Timer mechanism and window period, another reason that often leads to data inconsistency is the privacy threshold. 。 The privacy threshold is a “threshold” in this system, which means: only when enough people click or view the ad, the system will give the advertiser some data report. If there are too few people, the data will not be given, for fear of exposing personal privacy. For example: if you put an ad, only 5 people clicked it, the system will say: “Too few people, no data, fear of privacy leakage.” If 100 people clicked it, the system will give a summary report, such as “This ad brought 50 downloads”, but it will not tell you who clicked it. The privacy threshold of the specific channel determines whether CV can be returned. If an ad campaign is not installed enough, SKAN may not return CV data, but return NULL, resulting in missing data.
If the activity of the ad does not reach the privacy threshold set by Apple, the CV (conversion value) will become null, which means that Apple will not tell you what specific conversions this ad has brought, such as registration, purchase, adding to shopping cart, etc. Suppose you have placed an ad on a small channel, and only a few users clicked on the ad and installed the APP. Because the activity of this channel is too small, Apple believes that this data may involve privacy, and as a result, the conversion data is “hidden”. As a result, you cannot see what specific actions these users have taken, and there may even be no data returned at all. The money spent on this ad campaign is almost wasted. How to reduce the proportion of CV NULL? 1. Increase the advertising budget•Increase the budget, run more ads, and ensure that enough people see them. •For example: If your budget is too small and the ad exposure is low, increase the budget to let more people see the ad and increase the conversion volume. 2. Merge ad campaigns•Don’t split the audience too finely. You can merge multiple similar audiences to make the data volume of each ad campaign larger. •For example: You can merge audiences of different age groups to avoid only targeting 20-year-old users. 3. Take advantage of the new features of SKAN4.0 • Take advantage of Apple’s Postback function to obtain more data at different time points and reduce the risk of missing information. • For example: You can check the data multiple times within 24 hours to ensure that no important data is missed.
- Why does the data not match? 🍏 Apple’s black box mechanism + data delay Finally, let’s talk about why the data does not match. In fact, there are two main reasons for the data not matching: Apple’s black box mechanism and data delay. Apple’s black box mechanism SKAN’s attribution mechanism is controlled by Apple, and advertisers cannot directly obtain all attribution data. This leads to a problem: there may be differences between the data seen by advertisers and the data attributed by Apple. Advertisers may find that their data is more or less than the data attributed by Apple because of Apple’s black box mechanism. 🔒 Data delay Data delay is another reason for data mismatch. Due to the existence of the Timer mechanism and the window period, the data seen by advertisers is often later than the actual installation time. This leads to data delay. For example: suppose a user clicks on an ad and installs an app on January 1st, but due to the Timer mechanism and window period, Apple does not attribute this installation to the advertiser until January 30th. Then the data that the advertiser sees on January 30th will include the installation volume on January 1st. This will cause the data to not match.
- What to do if the data does not match? 🛠️Since data mismatch is inevitable, how should we deal with it? Here are a few suggestions for you: 1. Understand SKAN’s attribution mechanism First, advertisers need to understand SKAN’s attribution mechanism, especially the Timer mechanism, window period and privacy threshold. Only by understanding these mechanisms can we better explain the reasons for the mismatch of data. 2. Adjust the time range of data analysis Due to the existence of data delays, advertisers can adjust the time range appropriately when analyzing data. For example, don’t just look at the data of a certain day, but look at the data trend over a period of time. This can reduce the impact of data delays.