In the ever-evolving landscape of digital advertising, privacy concerns and regulatory changes have significantly impacted how advertisers target and measure their audiences. One of the most notable shifts came with Apple’s introduction of the AppTrackingTransparency (ATT) framework in April 2021, which drastically reduced the availability of user-level identifiers on iOS devices. In response, Verve Group developed ATOM (Anonymized Targeting on Mobile), a privacy-centric solution that has now evolved into its third iteration, ATOM 3.0. This article explores how ATOM 3.0 is revolutionizing mobile ad targeting by leveraging on-device and contextual signals to create audience cohorts without compromising user privacy.
The Genesis of ATOM
Responding to Apple’s ATT Framework
When Apple introduced the ATT framework, it created a seismic shift in the mobile advertising industry. The framework required apps to obtain user consent before tracking their activity across other apps and websites, leading to a significant drop in the availability of the Identifier for Advertisers (IDFA). This change forced app developers and their ad tech partners to seek alternative targeting solutions. Verve Group quickly responded by launching the beta version of ATOM in May 2021, which aimed to mitigate the impact of signal loss by using on-device and contextual signals to create audience cohorts.
The initial impact of the ATT framework was dramatic and widespread, leading many in the digital advertising industry to scramble for effective solutions to this new reality. Verve Group’s decision to introduce ATOM as a beta was timely, aimed to address the reduction in identifiable trackable data on iOS devices. ATOM’s approach of utilizing on-device and contextual signals represented a departure from relying solely on user identifiers, offering a new pathway for ad targeting that respected user privacy and adhered to Apple’s stringent data tracking guidelines.
Early Challenges and Adoption
The initial rollout of ATOM faced several challenges. Cohort-based targeting was a novel approach at the time, and publishers were hesitant to adopt the unproven technology. Additionally, training ATOM’s machine learning models to accurately classify users into cohorts required a substantial amount of data, creating a catch-22 situation. Verve Group needed more adoption to generate enough data to train the models, but they also needed well-trained models to drive adoption. Despite these hurdles, the company persevered, and ATOM began to gain traction in the industry.
Overcoming the skepticism among publishers and proving the efficacy of cohort-based targeting was a significant feat. Verve Group’s perseverance amid these challenges showcased their commitment to developing a robust and reliable ad targeting solution. As data accumulated and the machine learning models improved, confidence in ATOM grew. This gradually led to more publishers adopting the technology, facilitating a more data-rich environment that contributed to the enhancement of ATOM’s targeting precision.
The Evolution to ATOM 3.0
Technical Complexities and Integration
Developing ATOM 3.0 was not without its challenges. The integration of ATOM through an existing SDK, now bundled directly into the HyBid SDK (formerly known as PubNative), proved to be complex. However, this integration was advantageous for long-term adoption due to pre-existing app integrations. After overcoming several technical complexities and delays, Verve Group successfully launched ATOM 3.0, which is now fully integrated within the HyBid SDK, spanning over 10,000 app publishers and reaching more than 1.5 billion users.
The process of refining ATOM and ensuring its seamless integration into the HyBid SDK required meticulous efforts from Verve Group’s development team. The integration was strategically designed to leverage existing app integrations, thus facilitating a smoother transition for publishers. By embedding ATOM within the HyBid SDK, Verve Group could ensure broader adoption, enabling the tool to activate each time a user opens an app within their network. This approach not only streamlined ATOM’s operations but also enhanced its scalability across a vast network of users.
Enhanced Features and Capabilities
ATOM 3.0 brings several enhanced features and capabilities to the table. The tool activates automatically each time a user opens an app within the network, executing machine learning algorithms on the user’s device to analyze in-app contextual signals such as gestures, session length, and session depth. These signals are combined with device signals like keyboard language, battery level, and Wi-Fi connectivity to classify a user probabilistically into one or more cohorts. This on-device processing ensures that no deterministic user-level information leaves the device, maintaining user privacy.
The advancements in ATOM 3.0’s features underscore the commitment to user privacy while maximizing the effectiveness of ad targeting. By performing on-device processing, ATOM ensures that sensitive user data remains localized, enhancing privacy protections. This combination of various in-app and device signals allows ATOM to build accurate audience cohorts without compromising individual privacy. Such features not only improve ad relevance but also align with evolving privacy regulations, making ATOM 3.0 a forward-thinking solution in the digital advertising domain.
Privacy-Centric Approach
Respecting User Privacy
One of the key strengths of ATOM 3.0 is its privacy-centric approach. Unlike traditional targeting methods that rely on user-level identifiers such as IDFAs, mobile IDs, IP addresses, or encrypted emails, ATOM 3.0 only reads contextual signals within a specific time frame and app, with no cross-app data sharing. This approach aligns with the growing industry trend towards privacy-enhancing technologies (PET), driven by large platforms like Apple through ATT and the Android Privacy Sandbox.
The shift towards privacy-enhancing technologies represents a profound change in digital advertising, emphasizing the importance of respecting user consent and data sovereignty. ATOM 3.0’s model of contextual signal analysis within a single app session reflects this industry trend. By eliminating cross-app data sharing and eschewing traditional proxies like IDFAs, ATOM places user privacy at the core of its operations. This method exemplifies a responsible approach to data usage, catering to the heightened privacy expectations of modern users.
Broad and Granular Cohort Categories
Initially, Verve Group offers several broad cohort categories with more granular segments under each. Behavioral and interest-based cohorts include loyal or casual users, night owls, active gamers, sports fans, and individuals interested in a healthy lifestyle. Additional categories such as “personal finance” and “mobility” cohorts classify users by income level or whether they’re likely at home or work, respectively. ATOM can also infer demographic information like age range, gender, marital status, the presence of children in the household, and if someone is on a business trip.
The detailed nature of these cohort categories demonstrates ATOM 3.0’s capability to deliver highly targeted advertising without compromising user privacy. By leveraging a diverse range of signals to infer user characteristics and behaviors, ATOM can cater to specific advertising needs. This level of granularity provides advertisers with actionable insights, facilitating precise ad targeting. Moreover, the ability to infer demographic and behavioral information broadens the scope of targeted audience reach, making ATOM a versatile tool in the advertising arsenal.
Real-World Impact
Case Study: iFunny
For app developers like FunCorp, the publisher of iFunny, the advent of Apple’s ATT framework and the subsequent reduction in IDFA opt-in rates posed significant challenges. FunCorp’s Chief Revenue Officer Sergei Efimov revealed that iFunny’s revenue dipped noticeably post-ATT, with IDFA opt-in rates plummeting from 50% to 15%, later slightly increasing to 25% after several product modifications. Despite these efforts, the opt-in rate remained “insufficient,” leading iFunny to prioritize user acquisition on Android over iOS.
The reduced opt-in rates for IDFA on iOS significantly impacted iFunny’s revenue streams, compelling it to shift focus toward Android, where targeting capabilities remained less restricted. FunCorp’s experience encapsulates the broader industry struggle with the aftermath of the ATT framework. Even with subsequent modifications and strategies to bolster IDFA opt-in rates, the challenges in iOS user acquisition persisted. This situation emphasized the need for effective alternative targeting solutions to navigate the new privacy landscape.
Recovery Through ATOM
To mitigate these challenges, FunCorp partnered with Verve Group to integrate the ATOM solution, aiming to recover and improve their ad targeting capabilities. By leveraging ATOM’s privacy-centric approach, iFunny saw significant improvements in ad effectiveness and user engagement. The integration helped offset the losses caused by lower IDFA opt-in rates by utilizing sophisticated cohort-based targeting that respected user privacy. The positive impact of ATOM on iFunny’s ad revenue illustrated the potential of privacy-focused solutions in contemporary digital advertising, offering a promising pathway for other app developers facing similar challenges in the privacy-driven market landscape.
