Monday , May 17 2021

A complete guide for mobile device retailers. Part 2 – Marketing on

In this article, we will therefore continue to deal with the upcoming changes that Apple introduced to us this year in iOS14 (part one). In this article, we will answer all the remaining questions and add introductory information on changes to the allocation principle.

In the new iOS 14 essential the approach to allocation will change. If previously IDFA became the basis responsible for full transparency of user data transmission (including it was also responsible for the exchange of this data between analytics, advertising network, DMPs), now analytics and tracking platforms need to build their data based on fingerprints and the probability model to be built on data from SKAdNetwork reports.

Another important issue is delayed data creation. So, the time to transfer data on the implementation of certain actions in the application can now take up to 72 hours.

In general, the issue of data transfer using fingerprints and predictive modeling is not new for analytics systems, as previous installations with limited ad tracking (LAT installations) were considered to be using this model. Only that previously such a percentage of users was as small as possible. And this did not significantly distort the analytical data. Now this issue and the issue of building the forecast data is becoming a priority.

Another feature of iOS 14 – this lack of attribution of gaze… Now all the data is built on the information about the user’s interaction with the ad when they click on it. If we simultaneously analyze the indicators of new installations in different video networks, then a significant percentage of new users come only by assigning views. This is also worth considering when planning advertising campaigns in the full transition to the new allocation principle.

Now let’s talk about that SKAdNetwork.

Assigning with this method will allow you to transfer information with the following parameters: Publisher ID – Click source information Campaign / Ad Group / Ad Network ID – Information about your campaign / ad settings or some targeting-related features. Reload flag – user data (new or returned); Conversion value – 6-bit value that allows you to specify which flow event the user has performed. The default value is 0. Data cannot be downloaded in real time, nor can ROI, LTV, or deep connection data be included.

Re-targeting in iOS 14

Because we don’t have the ability to identify a user’s devices and their previous actions in the app as part of a new type of assignment, re-targeting in the “new mobile world” is impossible. Only campaigns for different products from the same publisher will be available for execution, where the chain can be closed with IDFV (seller identifier) ​​data. But it will basically redirect your audience to other products, which is basically similar to advertising in hyper-casual games, where traffic shifts from one product to another.

Frodo in iOS 14

Apple, for its part, uses cryptographic verification to verify attribution. However, in reality, this does not protect the advertiser from schemes that fraudsters are already in favor of, such as flooding clicks. Due to the lack of connection to the data transmitted via IDFA, anti-fraud systems cannot be sure of the involvement of a particular clique in fraud / nephrodis. Accordingly, the new allocation principle once again addresses the problem that popular MMPs and anti-fraud systems have long struggled with.

Contextual targeting

Targeting iOS 14 is now based on Word2Vec technology, a special type of machine learning. It allows you to decide on the placement of your ads based on the data in which applications a particular user is currently playing. This data allows you to analyze in detail the ASO and other parameters of the application, and you can decide whether to show ads in this application or not. Using this technology, we can analyze their behavior in real time without compromising their privacy. However, each of the advertising networks will combine this technology with its own internal development, which will make targeting even more targeted and in the real interests of users.

Ad frequency

Another problem with running ads without IDFA is limiting the frequency of ad impressions per user. Due to the lack of most data that allows ad networks to choose not to show ads to a particular user, the ad network will now build data based on the user’s location, device model, and usage time. This will allow you to form specific groups, depending on the size of which the frequency of advertising will be limited.

Working with the user interface / user interface of your product will help increase the proportion of users who complete IDFA

The main and important task of a mobile product is now communicate IDFA download value and mobile ad tracking capabilities.
What do you need to pay attention to?
1) Detailed determination of data on which systems and how much they use user data (we describe the name of tracking systems and ad networks through which customized ads will be displayed). Here are examples of how this is used on Tinder and TripAdvisor:

2) A detailed description of the data on why these systems use the data and an indication that this data does not affect the user’s privacy. We try to “calm down” the user, to say once again that we are maximally interested in the safety of each of our users. We attach examples of Instagram / Tinder:

Many products ignore the addition of an ad tracking section to a separate item in the app settings. However, by informing and explaining to users about the need to use this information, you increase the possibility that the user will allow the transfer of the advertising ID and will not restrict the tracking of ads on their device for your product.
Note that running the AppTrackingTransparency framework on iOS 14 will prompt the user once for a data collection notification. This notification is relatively standard, but its subtitle can be changed to describe to the user the information about what this information needs. It looks like this:

In fact, this notification has the greatest effect because it disables / enables the ability to restrict ad tracking. That is why it is the job of the marketer most concisely and clearly convince the user of the importance of sending this information. We recommend that you perform separate message tests when you receive information about which of the messages will show more efficient conversion to user consent for data transfer.

In addition, you can test exactly when to display this message. For example, if you display it after a user performs a specific action in the app, you will better understand the value of that product, which will ensure and strengthen a certain brand trust. It is also helpful to start exploring successful approaches to pushing notifications and consent forms. The example below is particularly good as it requires a formal push notification in front of the official consent form. Example:

You can also create separate forms of user interaction with a detailed explanation, referring to the detailed descriptions of the privacy policy and then again to the AppTrackingTransparency framework.

Apple Search Ads tracking changes

Apple Search Ads tracking has undergone some changes since February 15, 2021. In particular, the largest MMPs allow tracking and correct counting of user data coming from the ASA, independent of SKAdNetwork, allowing separate allocation of installations from the ASA and retrieval of the widest possible data and detailed analysis of user events, their revenue and LTV.

Note that only the latest version of the Analytical System SDK, which supports the AdSupport and iAd frameworks, is required for a correct calculation.

Note for proper operation with DSPs

If the DSP you are communicating with and working with has not yet assigned a unique identifier that passes its parameters to SKAdNetwork, follow the link to a guide on what needs to be done to do so. Once again, we recommend that you discuss your DSP in detail and share this guide with the Traffic Manager.

Source link