Since the release of Apple’s iOS 14 in 2020, through the adoption of iOS 14.5 in April 2021, and the most recent announcement of iOS 16, Adjust has been hard at work developing solutions and doing research on successful techniques. Our industry has shifted to embrace a new privacy-centric paradigm of data analysis and attribution, moving from best practices for obtaining the opt-in to assisting clients with the implementation of sophisticated SKAdNetwork (SKAN) and conversion value strategies.
This series examines opt-in and optimization strategies for raising consent rates. High ATT opt-in rates don’t just mean more consented data for granular analysis. The consented information you’re able to extract from devices that have provided access to the IDFA makes up essential building blocks for your broad SKAN strategy and for how you work with conversion values and predictive analytics.
Today, we’re taking a look at A/B testing with respect to the opt-in. For a quick review, we also suggest reading Design do’s and don’ts for website as well.
Since the release of Apple’s iOS 14 in 2020, through the adoption of iOS 14.5 in April 2021, and the most recent announcement of iOS 16, Trackier has been hard at work developing solutions and doing research on successful techniques. Our industry has shifted to embrace a new privacy-centric paradigm of data analysis and attribution, moving from best practices for obtaining the opt-in to assisting clients with the implementation of sophisticated SKAdNetwork (SKAN) and conversion value strategies.
This series examines opt-in and optimization strategies for raising consent rates. High ATT opt-in rates don’t just mean more consented data for granular analysis. The consented information you’re able to extract from devices that have provided access to the IDFA makes up essential building blocks for your broad SKAN strategy and for how you work with conversion values and predictive analytics.
Today, we’re taking a look at A/B testing with respect to the opt-in. For a quick review, we also suggest reading Design do’s and don’ts.
User Perceptions Of Data Privacy
Positively, research indicates that more consumers than previously believed are at ease with tailored advertising. One study performed by Oxford Economics suggests that around 70% of consumers are open to opting in for a tailored ad experience. Only a small minority, according to the data, as opposed to a more customized, targeted experience:
- Only 17% of people feel uneasy with tailored offerings.
- Only 15% of people find customised goods and services uncomfortable.
How comfortable consumers are with opting in is greatly influenced by their level of trust. Over 8,000 people were polled for another study by Salesforce to determine what trust meant to them. 75% of respondents chose “privacy,” and 70% chose “transparency,” demonstrating how closely trust and privacy are related as well as how clearly privacy regulations are conveyed.
The most important lesson here for mobile marketers is how crucial it is to convey the benefits of opting-in. This is the most important element that needs to be underlined because a large majority of customers are comfortable with tailored advertising. A tailored user experience with fewer but more pertinent adverts is what opting-in entails. The industry’s message around the value-add of opting-in appears to be improving, and consumer understanding of what the ATT prompt actually is is also on the rise, according to Trackier statistics on the increasing opt-in rates.
Suggested Methods For Privacy Notifications
Organizations generally use privacy notifications to outline how they handle personal data and how it relates to various data protection laws, such the GDPR. We conducted a review of industry standards to better understand how privacy notices are displayed on mobile devices. Three patterns in the typical presentation of information were observed:
- Some applications provide users with complete data control and the most detailed opt-in choices.
- Some apps restrict control by not disclosing all of their collaborators.
- A third category of apps employs a “all in” or “all out” strategy, allowing users to choose whether they want to share their data with each and every one of the partners specified or not.
We will always advise companies to follow the first strategy: be as open and give users complete control over their data as you can. To bolster this, you can provide convincing content and images that explain why data gathering is necessary and list the advantages of opting-in.
Grouping
Many businesses questioned whether it would be possible to serve the ATT prompt request with other privacy notices when iOS 14 was initially released. While a pre-permission prompt is fully Trackable, only the second string of the ATT prompt is, and no extra notices may be added. It’s also important to keep in mind that users cannot be pressured into responding by having opt-in checkboxes already selected or by having an opt-in CTA.
According to one study, providing users with two ways to provide consent and phrasing the message positively increases the likelihood that they will opt-in, as seen in the example below. It’s crucial to keep this kind of positive framing in mind while discussing privacy notices in general. You should accentuate the advantages.
A/B Testing
By contrasting two opt-in tactics and gauging their effectiveness, A/B testing is an excellent way to examine your solution.
We advise you to do an A/B test before showing your opt-in message as a solo message or bundling it with the GDPR privacy notice at pre-permission levels. Don’t forget to imitate Apple’s ATT pop-up once a user accepts your opt-in message.
Below, we’ve detailed test rounds with several components to aid you in developing a research strategy.
Then, by adding other factors, you can build on these findings. For instance, if a privacy notice bundle that include the Apple request is more effective, you might investigate the impact of various language or design choices on opt-in rates. As an alternative, you can evaluate the timing of when it’s served if you discover that displaying a solitary request (such as a pre-permission prompt or the Apple pop-up) was more successful.
You might also think about utilising log linear analysis to assess the effects of multiple variables on opt-in rates if you have a sizable user base and sufficient resources. Additionally, we advise assessing how frequently you provide your opt-in strategy to consumers who didn’t initially opt-in.
You can investigate whether certain user segments have statistically significant effects. It’s possible that users from one location opt-out more frequently than users from another, or that opt-in rates for new users are higher than for existing users. With this information, you may move a step closer to dynamically changing your approach to raise opt-in rates.
To interpret the results of any A/B testing, you should compute a confidence interval. This aids in determining the range in which the actual opt-in rate would fall if the test were run with all of the users of your app.
Predictive Modelling To Forecast The Opt-in
To forecast specific user actions, predictive modelling use statistical techniques. You can use two different kinds to analyse your A/B tests:
The link between variables is examined through regression analysis. Based on predictor factors, it can be used to forecast the value of an outcome variable.
Based on the observations of the input variables, a target variable’s outcome is predicted using decision tree analysis.
Decision tree analysis and logistic regression are both effective approaches for categorization issues. If you think that your data set separates linearly into two parts, one part connected with the decision to opt-in and the other part linked with the decision to opt-out, logistic regression is usually the superior method. Regression analysis should also be used if your predictor variables’ values are continuous.
A decision tree, however, is a better fit if you’re not sure about the data separation. Additionally, a decision tree is frequently the best option if your dataset has a high percentage of outliers, missing values, or is skewed.
We advise you to try both approaches first before deciding which model produces the greatest results. The next stage is to evaluate each predictor variable’s unique contribution to determine which factors—such as install type, geography, demographics, etc.—have the most impact on user decisions.
Speak With Your Users To Learn About The Motives Of Your Customers.
A/B testing and regression analysis will reveal which elements are most likely to boost user opt-in rates, but these tools won’t explain why the strategy is effective or why some variables are more crucial than others. This is ultimately accomplished through talking to your users and performing in-depth interviews that translate quantitative results into clearly defined decision-making processes. You may further refine your dynamic opt-in strategy by using the information from these interviews, both with people who have opted in or are likely to do so and with those who have opted out.
Summary
A revolutionary topic in the realm of mobile and app marketing is data and user privacy. At Trackier, we firmly think that every business should put this at the heart of how they handle data and interact with customers. In the end, adopting a clear and transparent strategy will assist develop user trust and increase their willingness to opt-in to providing their IDFA.
We are now seeing a moderate and steady rise in overall opt-in rates, nearly 18 months after the release of iOS 14.5 on mobile devices. You’ll steadily improve your chances of establishing trust with your users and getting high opt-in rates by testing and determining how to optimise your consent flow. For additional information, get in touch with your Trackier contact!