Expert insights from Kochava webinar with Spotify Advertising and M&C Saatchi Performance

Since its emergence in the early 2000s, podcasting has experienced exponential growth, prompting businesses to adopt marketing strategies to leverage the rapidly evolving medium. In a webinar showcasing podcast advertising, acclaimed industry players—Charles Manning, CEO of Kochava; Adrienne Rice, Director of Media Investment at M&C Saatchi Performance; and Matt Drengler, Director of Marketing Research and Intelligence for Spotify Advertising—shared their insights into the opportunities and best practices within this channel. The insightful session examined the breakthroughs of podcast advertising, its efficacy for advertisers, and its future.

Why Podcast Advertising?

The panel established the impressive scale of the podcasting landscape, emphasizing the medium’s growth and the opportunities this affords advertisers. Millions of podcasts cater to a global listener base projected to exceed half a billion people in 2024. At the same time, podcast advertising revenue is heading upward of $4 billion. This is no longer a nascent, niche medium, but a burgeoning channel with yet-untapped potential.

Rice shared key insights into the demographics and behavior patterns of podcast listeners that marketers might do well to consider, pointing out that 66% of US internet users listen to podcasts (in most cases at least once a week), with the majority of listeners aged 45 or younger and earning higher-than-average household income. As for the dominant podcast genres—spoiler alert, but perhaps not a surprise—comedy and true crime are well established as top listener favorites.

Evolution of Podcast Advertising

The panel recounted the transformative journey of podcast advertising, from its early implementation to the innovative solutions shaping its future. Traditionally, podcast advertising was predominantly purchased directly from shows and embedded within the podcasts themselves, often in the form of host-read ads. This posed significant limitations in scalability, as each ad insertion was manually placed within and inseparable from the episode. Producers and advertisers also had to consider the ongoing relevance and permanence of ad content; after the show was aired, the ad might be forever “baked in.”

The podcast industry embraced technological innovations to solve these challenges, and over the past several years, the landscape has significantly evolved. Drengler took participants through industry advancements that directly addressed the limitations of embedded ads and revolutionized the podcast advertising space, most notably dynamic ad insertion (DAI). This technology, now accounting for 90% of ad volume, enables advertisers to place relevant ads into a designated spot within a desired podcast episode, seamlessly stitched in at the time of download and refreshable as needed. This marked a significant advance toward resolving issues such as scalability, measurability, and systematic targeting.

Automated programmatic ad placement is rapidly taking hold, and there is still much room for growth in this approach. And Spotify’s streaming audio insertion (SAI) represents a cutting-edge breakthrough, leveraging the shift toward streaming podcast content rather than downloading it. This technology has further enhanced ad integration, real-time targeting, dynamic content delivery, and ad measurement capabilities, in particular the ability to measure on real-time impressions, leading to a more engaging ad experience for listeners while offering greater effectiveness and efficiency in optimization for advertisers.

New Possibilities

Manning emphasized that these significant shifts now enable us to comprehend the podcast consumer journey holistically, essentially having blown open the doors for the medium to become fully viable for performance advertising. The webinar panel agreed that new technologies are rapidly driving equivalence with digital ad formats, fully democratizing podcasting as a reliable advertising channel.

Taking full advantage of these advancements, however, necessitates new paradigms in measurement. Spotify’s SAI offers advertisers a more precise measure of reach, impressions, and audience targeting. While this allows for sophisticated metrics, the greater podcast adtech world is still catching up. Case in point—in a digital environment where clicks and downloads are often misleading, distinguishing between podcast downloads and streams is key to tracing listeners’ post-impression actions.

To facilitate such measurement capabilities, Spotify partnered with Kochava to process and analyze a more dimensional profile of podcast stream data in real time. Advertisers on this platform are no longer subject to the limitations posed by engagement ambiguity as revealed solely by tracking downloads or one-touch attribution. The Spotify-Kochava collaboration has enabled third-party verified measurements that open the way for further performance-based initiatives. One actionable metric has revealed that up to 95% of attributed events take place within 14 days of podcast download or exposure.

Effective Campaigns and Best Practices

These insights derived from enhanced measurability reinforce the importance of understanding the customer journey and the role of podcasts in this journey, from introduction to final conversions. Podcast advertising is more than just another channel, but a uniquely immersive experience that provides a focused and uninterrupted space for advertisers. The conversation revealed a bombshell outcome takeaway: One in five listeners who visit an advertiser’s site after exposure to a podcast ad ends up making a purchase. Ponder that!

The panel delineated key practices for devising and executing successful podcast campaigns:

Leverage listeners’ heightened attention: Advertisers need to comprehend the medium’s perceived authenticity and credibility for effective education and audience engagement over a wide range of topics, resulting in a loyal, receptive listener base. The felt connection between host and listener fosters trust in the medium and by extension the advertisers who directly speak to this audience engagement. High-quality, vivid creative is a must to engage podcast listeners who are primed to embrace relevant, compelling ads and brands/products that complement their listening experience.

Deploy a robust measurement strategy: Advertisers need to leverage the wealth of data now available through podcast analytics. Understanding listener behavior, such as when and how they tune in, listen to or skip ads, and engage with content, is fundamental for optimizing campaign performance. Contextual-based targeting, including seamless, real-time topic and conversation-specific ad placements, is a powerful means to tailor creative to podcast contexts and/or home in on audiences by demographic or behaviors and interests. Data derived from such practices can be used to inform and optimize subsequent initiatives relative to desired key performance indicators.

Prioritize privacy issues: With privacy becoming an increasingly important concern, advertisers need to be cognizant of how they collect and use listener data. Ensuring compliance with privacy laws and being transparent with listeners about data usage can help maintain trust and reinforce positive brand image.

Catch the Full Webinar on Demand

The complete on-demand webinar, Capitalizing on Podcast Advertising in 2024, is available now! The discussion is full of fascinating insights on podcast advertising, effective measurement approaches, and future trends, with a fun addition of some of the speakers’ own favorite podcasts. The overall takeaway from this informed panel of industry experts: It is abundantly clear that the podcast medium will continue its upward trajectory, and savvy marketers will be eager to leverage this golden opportunity to apply these webinar insights directly into their digital marketing strategies for a marked competitive edge.

Answers to your questions from the Kochava Foundry webinar

Grant Simmons, VP of Kochava Foundry, and Ethan Lewis, Chief Technology Officer at Kochava, recently hosted the webinar Navigating Google Privacy Sandbox—Part 1, where they unpacked the industry’s upcoming sea change with Google’s rollout of Privacy Sandbox for Android and spotlighted key trends in the shift in mobile toward user privacy. In this follow-up, they have compiled audience questions to address and elaborate upon in further detail.

Check out the full webinar on demand here.

#1 Has Google published the timeline for deprecating their Advertising ID (ADID) from Android?

Google has not yet published a definitive timeline for the deprecation of ADID, also sometimes called Google Advertising ID (GAID), from Android. The ADID/GAID is anticipated to follow a similar path as Apple’s IDFA insofar as its utility for tracking and measurement is expected to diminish. Given Google’s significant stake in the adtech ecosystem, their ADID phaseout may be more gradual compared to Apple’s rapid deprecation of IDFA. There are indications the deprecation process could begin with the phasing out of third-party cookies, expected to start this fall. Google’s active development of APIs for Privacy Sandbox signals a move toward testing with publishers later this year, with a broader rollout and ADID deprecation potentially starting next year. Marketers should prepare for a future where public unique identifiers such as ADID are no longer available and seek alternative privacy-centric measurement solutions.

#2 Is Google going to deprecate Google Play Install Referrer?

While Google has not made a formal announcement regarding this, there are indications they may deprecate the use of UTM parameters, which are critical for mobile tracking as they can be picked up via the Play Store and used to power Google Analytics. The potential deprecation of these links could begin next year, signifying a pivotal shift in mobile tracking and analytics.

#3 How does this compare to iOS App Store data restrictions?

Google’s Privacy Sandbox and Apple’s SKAdNetwork (SKAN) share the goal of enhancing user privacy while providing campaign performance metrics. Both are designed to be anonymous while offering event-level reporting. Their approaches differ, however, with Privacy Sandbox developed through broader community collaboration, while SKAN is an Apple-led initiative. Privacy Sandbox aims to provide tools for targeted advertising without individual user tracking, whereas SKAN offers a more limited framework for iOS app advertising attribution. Advertisers face challenges with both due to reduced granularity of data.

#4 How does this impact MMM, if at all?

Marketing mix modeling (MMM) is likely to thrive, as it relies on modeling of aggregated data as opposed to the granular data necessary for last-touch attribution. MMM platforms, such as AIM by Kochava, can ingest SKAN and Privacy Sandbox data to power their models and help marketers understand influence and incrementality across channel partners. Separately, mobile measurement partners (MMPs) will play a crucial role in understanding data connections, providing tailored measurement solutions, and syndicating measurement data as needed.

#5 How will Privacy Sandbox impact app remarketing, both gaming and non-gaming?

The exact mechanisms for user suppression or retargeting within Privacy Sandbox are not yet clear. However, it is expected that aggregate data will be managed via API, with flags indicating prior customers vs. new ones. Brands will need to differentiate between new and existing customers and communicate this information to the networks they engage with for remarketing. They should also continue to invest in owned media as a pillar of their remarketing strategy.

#6 How will Privacy Sandbox work for user acquisition? How is Kochava thinking about its role working with SDK-less partners, the delegation functionality in PAAPI, and PAS?

Google Privacy Sandbox is set to introduce new frameworks for user acquisition that prioritize user privacy. For instance, the Attribution Reporting API within Privacy Sandbox will enable advertisers to measure campaign performance without relying on traditional identifiers. As an MMP, Kochava is preparing to adapt to these changes by exploring SDK-less integrations and server-to-server clean room integrations. Kochava—an approved testing partner with Google—is actively involved in testing these new mechanisms. The second part of our Google Privacy Sandbox webinar series will delve deeper into how these integrations will function as well as the role of Kochava in this evolving landscape. It will also address to what extent Kochava will interact with the Protected Audiences API (PAAPI) and Protected App Signals (PAS).

#7 How will cookie deprecation impact DSPs and SSPs since they heavily rely on pixels? Do we know what Privacy Sandbox for app tracking will look like? What do we know of the differentiators as compared to SKAN?

Cookie deprecation will significantly impact DSPs and SSPs that have traditionally relied on pixels and third-party cookies for targeting and tracking. With Privacy Sandbox, Google aims to replace these methods with privacy-first alternatives, such as the Topics API for interest-based advertising and Attribution Reporting API for campaign measurement. These changes will challenge DSPs and SSPs to adapt their strategies, possibly leading to increased use of data clean rooms and data lakes. Google’s Privacy Sandbox for app tracking is expected to share similarities with Apple’s SKAdNetwork (SKAN), such as privacy-enhancing technology and anonymous reporting, albeit with its own unique approach to rollout, collaboration, and distribution effects.

#8 Is managing Google Privacy Sandbox on the roadmap for Kochava?

Kochava is an authorized testing partner with Google and actively engaged in managing the transition to Privacy Sandbox. The company is testing the new APIs and frameworks to assess their implications for mobile attribution and develop solutions that align with the privacy-first direction of the industry. As part of their commitment to adapting to these changes, Kochava will be integrating Privacy Sandbox features into services to help clients navigate the new landscape, with a strong initial focus on the Attribution Reporting API.

#9 Is Google Privacy Sandbox going to cost anything for the agencies that use it?

While there may not be direct costs associated with using Privacy Sandbox, the shift to privacy-first attribution methods will require agencies to adapt their strategies and potentially invest in new technologies or partnerships. The changes brought by Privacy Sandbox will be integrated into the adtech ecosystem, and agencies will need to evolve their practices accordingly. This evolution may involve indirect costs related to training, technology adoption, and changes in campaign management.

#10 What is the biggest challenge with Google Privacy Sandbox, and is there an upside of Google Privacy Sandbox from a marketing standpoint?

The biggest challenge with Privacy Sandbox is the shift away from deterministic attribution methods, requiring marketers to adopt more aggregated and model-based approaches to measurement. For the marketing industry, this will demand a new mindset and potentially new skill sets. On the other hand, the upside is an increased focus on consumer privacy, which may enhance trust and potentially improve the public perception of the advertising industry. Marketers will need to become more creative and strategic in how they target and measure campaigns, focusing on privacy-preserving methods that align with consumer expectations.

#11 Is there a POV on retention analytics and how this is going to be impacted/go away?

Retention analytics in the context of Privacy Sandbox remains an area of uncertainty. However, it is expected that technology solutions will be developed to assist with this aspect of analytics. Google has demonstrated a collaborative approach in the development of Privacy Sandbox, which suggests that feedback from stakeholders will influence shaping the future of retention analytics. It is important for marketers to stay informed and adapt to new tools and methodologies that emerge as Privacy Sandbox evolves.

#12 How does identity work in Privacy Sandbox for Android? Is it still based on advertising identifiers?

In Privacy Sandbox for mobile, identity will not rely on publicly available unique advertising identifiers. Instead, Google will utilize aggregated and anonymized data based on user information associated with Google accounts. This approach aims to preserve user privacy while still providing useful data for advertisers. The data will be structured to prevent the identification of individual users, aligning with the privacy-first initiatives of Privacy Sandbox.

#13 As a user, will I be able to opt out of certain interest topics within the Topics API?

While it is unclear whether users will have the ability to opt in or out of specific topics within Privacy Sandbox, it is expected that a new consent mechanism will be introduced on Android, similar to Apple’s App Tracking Transparency (ATT) framework on iOS. This mechanism will likely govern user consent for data collection and use in a privacy-conscious manner.

#14 What about gaming in the Topics API? Will it be broken down by subcategories?

The granularity of the Topics API, particularly for gaming, is not yet fully known. Initially, it is expected that categories may be broad and not provide the level of detail desired by performance marketers in the gaming sector. As Privacy Sandbox matures, however, it is possible that more specific subcategories would be introduced. In the meantime, marketers should focus on leveraging Event and Summary API data, which may offer more actionable insights in the early stages of Privacy Sandbox implementation.

#15 DSPs have spent a lot of time building out high-performance targeting products, but with Privacy Sandbox, they have to work within the browser or on device. How handicapped will their technical capabilities be if they can’t host massive amounts of campaign/targeting data in the browser memory? Or can they?

Demand-side platforms (DSPs) will face significant challenges as they adapt to the constraints of Privacy Sandbox, particularly with its limitations on using browser or on-device storage for campaign and targeting data. The extent to which DSPs can utilize such storage is uncertain, and it is likely that such capabilities will be restricted to ensure user privacy. DSPs may need to explore alternative strategies to comply with the new privacy regulations, relying less on extensive data storage within the browser.

#16 Will event-level reporting postbacks in Google Privacy Sandbox for Android have any kind of delay as with Apple’s SKAdNetwork?

Event-level reporting postbacks within Privacy Sandbox will indeed include delays similar to those in SKAdNetwork. These delays are part of the privacy-preserving features designed to prevent identification of individual users. The specific mechanisms and timing of these delays may differ from those in SKAN, and we expect to be able to clarify further details in the second part of our Google Privacy Sandbox webinar series. Marketers should anticipate adjustments to their reporting and analysis processes to accommodate these delays.

Got more questions on Google Privacy Sandbox?

If you seek clarity on how Google Privacy Sandbox for Android will impact your mobile marketing strategies or have specific concerns about this landmark transition, Kochava Foundry is ready to assist. Our team of experts can provide guidance on navigating these changes and help you adapt your mobile app campaign strategies for success in a privacy-first landscape. Set up an expert consultation with us to explore how we can support your needs and keep you ahead in the evolving digital advertising ecosystem.

Enhancements to MMM make it a powerful tool for advertisers as privacy regulations evolve

What Does MMM Stand For?

Known as marketing mix modeling or media mix modeling, MMM is a statistical data-analysis methodology that gives marketers a better understanding of the optimal mix of marketing strategies across media channels to positively impact sales and other key performance indicators (KPIs). MMM seeks to take into account all advertising channels—print, social, and online advertising (e.g., search, display, video) as well as offline channels.

How Does MMM Work?

Marketing mix modeling relies on aggregate data from marketing and non-marketing sources gathered over an extended period of time. Typically, a minimum of three or more months of historical data (ideally 12+ months) are necessary to reach data significance and account for seasonality shifts. This large volume of data is used to create an accurate demand model to give marketers insights into the most effective channel strategies and forecast the best omni-channel allocation of future ad spend for greatest impact and return on investment (ROI). From these insights, marketers can adjust ad spend allocation across channels and partners for future optimization.

Seems a little abstract? Let’s look at an example.

A CMO at a major fintech company wants to zoom out from granular campaign- and creative-level performance reporting to capture the bigger picture. The CMO’s goal is to understand the incrementality of ad spend across channels and overarching performance peaks and valleys throughout the year. They work with an MMM platform and after plugging in historical data are able to arrive at new recommendations for how much to spend across channel partners at different times of the year. The marketing director and UA manager now can reallocate spend across their various channel partners to drive better incrementality and reduce unwanted oversaturation on any given channel.

And that’s essentially MMM—collecting and processing a lot of data, then presenting it at a high, aggregated level so marketers can glean broad insights into advertising effectiveness, transcending individual outliers and skewed averages.

As digital marketing evolved in the early 2000s, MMM largely took a back seat to direct-response attribution modeling, which relies on user-level interactions on websites and mobile apps. Unlike MMM aggregated data, attribution data is inherently granular—useful for marketers in focusing their efforts on specific users and customers via direct response marketing. This approach facilitates insights derived from customer-level engagements, enabling marketers to drive creative optimization, A/B tests on messaging and creatives, and other personalized marketing tactics tailored toward unique persona profiles.

In recent years, however, MMM has seen a renaissance owing to the data processing and analysis potency enabled by AI and machine learning. Companies and their marketing teams have adopted the advanced analytics and predictive insights afforded by MMM to fuel growth. At the same time, recent developments in user privacy and data use have eroded the availability of granular, user-level attribution data. As a result, marketers are relying more on aggregated data and rediscovering the potential of MMM to inform their marketing strategy. MMM enables them to optimize budgets across channels while respecting privacy policies.

How MMM Is Evolving to Help Advertisers

With the revival of marketing mix modeling, how marketers interact with it has evolved to support the dynamic needs of today’s user acquisition teams. In the fast-paced digital advertising landscape, quarterly or semiannual MMM reports are quickly outdated and lack actionability. Traditional MMM is time-consuming and laborious to manage, making it accessible only to large organizations that have the resources to maintain it in-house or the budget to outsource it.

While historically only such corporations have been able to afford fully leveraging MMM, automated data flows, cloud computing, and machine learning have made MMM more accessible, accurate, nimble, and easily updated. Cutting-edge software as a service (SaaS) next-generation MMM solutions, now accessible to companies of all sizes, have been developed to fit the needs of today’s advertisers. AIM (Always-On Incremental Measurement) by Kochava, a real-time MMM tool, maximizes the effectiveness of the marketer’s budget by providing advanced control over incrementality, channel saturation, and seasonality. AIM utilizes a sophisticated learning system that ingests new data daily and continuously updates and enriches its models. This always-on approach ensures that the insights it produces never go stale and are always ready to use—providing marketers who must make confident decisions with turnkey recommendations for optimized budget allocations.

As user privacy continues to weave itself throughout the adtech ecosystem, next-generation MMM tools will become increasingly indispensable for advertisers in determining the effectiveness of their omni-channel media strategy.

The Conclusion on Marketing Mix Modeling

Next-generation MMM is at the forefront of a marketing revolution, offering actionable recommendations for data-driven decision making in an increasingly privacy-conscious adtech landscape.

Have questions or want more information on AIM and MMM? Check out our Marketing Mix Modeling 101 ebook and explore even more helpful content in the AIM Resource Center.

Subscribe to our newsletter to stay up to date on industry trends.

How to become an early testing partner with Kochava

Google Privacy Sandbox for Web has recently come increasingly under the microscope as the adtech industry witnesses early signs of third-party cookie deprecation’s impact on ad monetization across the open web. With Google’s 1% third-party cookie deprecation beta for Chrome users starting in early January, initial observations have noted Chrome users without cookies monetizing approximately 30% worse than users with cookies.

Propelling the adtech industry toward a more privacy-first approach is a massive undertaking, especially for the most dominant mobile and browser ecosystem in the world. Google is taking a collaborative approach with the industry to tackle this monumental shift, and Kochava is thrilled to be partnering with industry leaders such as IAB Tech Lab to ensure that Privacy Sandbox meets our customer’s needs. As a longstanding mobile measurement partner (MMP), Kochava is particularly focused on the coming of Privacy Sandbox for Android—and its implications for the mobile ecosystem.

A Refresh on Privacy Sandbox for Android

In August 2019, Google launched Privacy Sandbox as an initiative to develop new standards for websites to access Chrome user information without compromising user privacy. In February 2022, Google announced that Privacy Sandbox would be coming to its mobile operating system, Android. Privacy Sandbox for Android is often likened to Apple’s SKAdNetwork (SKAN), a privacy-enhancing technology for understanding iOS campaign performance in a privacy-first world, although the scope and impact of Sandbox will extend beyond SKAN’s purview.

In their own words, here are Google’s stated goals with Sandbox for Android:

So what are the tools in the Sandbox? As illustrated in the following graphic, Privacy Sandbox on Android consists of four primary technologies. Let’s unpack each in further detail.

Attribution Reporting API

The Attribution Reporting API serves as a privacy-first solution for marketers to measure the effectiveness of their advertising campaigns. It facilitates the aggregation of conversion reporting data (triggers) from different sources (i.e., attribution data from an ad click or impression) while maintaining individual user privacy. Using this API, marketers can assess the impact of campaigns without compromising individual user identities—ensuring privacy compliance while still providing a base level of performance insight for the purposes of campaign optimization.

Similar to SKAN for iOS, the Attribution Reporting API within Sandbox features privacy-preserving thresholds and outputs only anonymous, aggregated performance data. No user or device-level data is available. Unlike SKAN, which originally supported only app-to-app conversion paths (until the release of web-to-app support for Safari in SKAN 4.0), Sandbox will support app-to-app, app-to-web, web-to-app, and web-to-web user paths from the outset.

This API supports observance of measurement data through two types of attribution reports:

• Event Level Reports connect specific attribution sources from an ad click or ad impression with trigger data from conversions. The fidelity of signal output is more limited, as the connection is one-to-one.
• Aggregatable Reports provide a richer fidelity of trigger conversion data, but in only an aggregate format not necessarily tied to particular attribution source data.

Kochava is currently focused on testing Event Level Reports, which more closely resemble the style of reporting through SKAN on iOS.

Why it’s important
The current state of mobile attribution on Android relies on Google Advertising ID (GAID), UTM referrer, and/or other device characteristics, including user agent and IP address, that may be transmitted off device to perform one-to-one attribution between an ad impression or click and the resulting conversion. The Attribution Reporting API will eliminate reliance on this user and device-level data and bring advertising measurement on device. Sensitive signals will no longer need to be sent off device—making them unavailable for unauthorized collection, use, and covert tracking. With the eventual deprecation of GAID, UTM referrer, and access to other device signals, the Attribution Reporting API will be the lifeline through which marketers can understand the performance of their campaigns to inform their optimization decisions.

See Google Developer Documentation HERE.

Protected Audience API (formerly FLEDGE)

Originally named FLEDGE, now affectionately called PAAPI (Protected Audience Application Programming Interface), this set of APIs aims to support on-device auctions for remarketing and custom audience segmentation based on interest groups. The goal is to serve personalized ads to users in line with previous app engagement, but without any third-party data sharing.

Why it’s important
User data no longer needs to be sent off device for the purposes of building user profiles attached to GAIDs or other device/user-data derived profiles for personalized ad targeting across ad networks, DSPs, and other ad platforms. Adtech vendors will be able to tap into anonymous, yet highly accurate signals to inform ad buys based on user behaviors, interests, and historical app usage.

See Google Developer Documentation HERE.

Topics API

The Topics API in Google’s Privacy Sandbox for Android is designed to give marketers a privacy-centric method to target relevant audiences based on their interests. Advertisers can understand the topics engaged by users and serve them personalized and targeted ads without revealing individual user identities—respecting user privacy and maintaining data confidentiality. A topics taxonomy will provide hundreds to potentially thousands of human-curated interest labels that help categorize a user by interests.

Why it’s important
One might liken this to IAB Tech Lab’s Audience Taxonomy, which provides standard nomenclature for the classification of audience segments. The Topics API will provide the new standard for classifying Android users for targeting purposes by leveraging on-device learning. This replaces ad tech platforms collecting user and device data to build their own profiles on users attached to GAIDs or other third-party generated identifiers.

See Google Developer Documentation HERE.

SDK Runtime

SDK Runtime establishes a more secure framework for apps integrating third-party software development kits (SDKs). Because app developers are not always aware of a third-party SDK’s full functionality and data collection practices, SDK Runtime places third-party SDKs into a modified execution environment featuring well-defined permissions and data access rights privileges.

Why it’s important
Over the years, adtech news publications have featured many stories about rogue, third-party SDKs behind advertising fraud schemes, covert data collection, and other nefarious practices. While these SDKs were intended to leverage valuable app functions and features, rogue actors have been known to hide covert functionality deep within their codebase, enabling them to exploit data-access permissions for nefarious purposes, unbeknownst to the developer who integrated them for legitimate use cases. SDK Runtime technology will put third-party SDKs in a dedicated runtime environment that makes such exploitation impossible—giving app developers and the end consumer peace of mind.

The complete library of Kochava Android SDKs will be available through SDK Runtime.

See Google Developer Documentation HERE.

MMPs and the Attribution Reporting API

Let’s zoom in on the Attribution Reporting API—a key focus for the team here at Kochava.

Mobile measurement partners (MMPs) are able to integrate with the API to provide conversion analytics and performance insights for advertisers under the new privacy framework of Sandbox. It’s important to note that while ad network vendors can use the API to receive self-attributed event and summary reports for conversions they drive/influence, only an MMP is positioned to provide cross-network, last-touch attribution by integrating with the array of aggregation services set up by various ad network vendors. Google lays out multiple scenarios for cross-network attribution with an MMP in this developer documentation. Similar to how MMPs work as a unified decoder ring of sorts for the various SKAN-enabled media partners with which a brand is running campaigns, MMPs will again be sitting at the intersection, translating cross-network Sandbox data into a holistic reporting layer marketers can make sense of.

The Attribution Reporting API also provides for lookback window configurability adjustable by the advertiser and/or via their MMP partner. This is more flexibility than we see on SKAN, where such windows are fixed. Sandbox also provides 30 days of post-install event measurement for better user quality and retention insights out of the gate, compared to what SKAN offered at launch.

As neutral third-party measurement services, Kochava and other MMPs play an important role in the advertising ecosystem. The Attribution Reporting API provides both event-level and aggregated attribution reporting to MMPs, which along with other aggregated omni-channel data helps MMPs empower marketers to understand overall campaign effectiveness and optimize spend across multiple media channels. The Privacy Sandbox model creates opportunities for MMPs to innovate with privacy-focused solutions that decomplicate the lives of marketers amid the increasingly complex privacy considerations of digital advertising.

Kochava Sandbox Testing

Kochava engineering and Android SDK development teams have commenced testing of the primary Attribution Reporting API flow:

1. Registering ad clicks or views (impressions) that lead users to a particular app or website to complete a conversion (known as attribution sources)
2. Next, registering triggers (conversions) that signify a user taking a valuable action such as installing an app, making a purchase, or starting a free trial
3. The Attribution Reporting API receiving both attribution sources and triggers, making relevant matches for conversion attribution and sending one or more triggers off device through event-level and aggregatable reports

Are you interested in Sandbox testing with Kochava?

While testing is already underway with a small selection of clients and partners, we’re looking to expand our testing group. Please note that currently our testing is focused on Event Level Reports.

Stay Updated

Privacy Sandbox for Android is a multi-year effort, and Google has not given an exact timeline for general release. Subscribe to our newsletter to stay connected and up to date on future Privacy Sandbox milestones and related updates to Kochava products and services. You can also enroll for notifications directly from Google.

Answers to your questions from the Kochava Foundry webinar

Grant Simmons, VP of Kochava Foundry, recently hosted the webinar Measuring Incrementality & Lift, where he unpacked lift measurement best practices, common pitfalls with hold-out groups, emerging methodologies for incrementality testing, and Kochava solutions for incremental lift measurement. The Foundry team gathered some of the most engaging audience questions Grant answered during the webinar to elaborate on in further detail.

Check out the full webinar on demand here.

How exactly does Kochava Foundry measure incrementality & lift for content partners that run ads year round with no dark periods?

Foundry employs various measurement techniques to assess incrementality and lift for content partners with ongoing, year-round ad campaigns. While MediaLift is primarily designed for discrete campaign measurement, Foundry can utilize modeling techniques to understand the ongoing lift of always-on media.

By leveraging regression discontinuity, Foundry can analyze the impact of continuous advertising efforts and identify the incremental contribution of specific media partners. Additionally, AIM (Always-On Incremental Measurement) can be employed to measure the contribution of each channel continuously at the network, publisher, and campaign level. This enables content partners to gain insights into the effectiveness of their ongoing ad campaigns and make data-driven decisions to optimize their media mix.

Can Kochava MediaLift be used for an advertiser that is not using Kochava as their mobile measurement partner (MMP)?

Yes, Kochava MediaLift can be utilized by advertisers who don’t have Kochava as their MMP. MediaLift is designed to work with standardized ad signals and conversion signals, making it platform-agnostic. This means that even if an advertiser uses a different MMP, MediaLift can still analyze the ad and conversion data to measure incrementality and lift. By leveraging MediaLift, advertisers can gain valuable insights into the incremental impact of their campaigns, regardless of their chosen MMP.

What are some best practices for running incrementality tests while minimizing negative impact to my business?

• Define clear objectives: Clearly define what you want to measure and the specific metrics you are targeting. This guides the design of your test and ensures that you are capturing meaningful data.
• Consider opportunity and hard costs: Holdouts, or not marketing to a portion of your audience, can be costly and result in missed opportunities. Consider the potential impact on your business and weigh the costs against the benefits of the test.
• Determine timing and duration: Choose the appropriate timing and duration for your test to capture meaningful data without disrupting your ongoing campaigns. Consider factors such as seasonality, campaign duration, and audience behavior to ensure accurate results.
• Explore modeled options: Modeled approaches, such as synthetic control groups or machine learning models, can provide reliable results while mitigating the opportunity cost of holdouts. These approaches use historical data and statistical modeling to estimate the incremental impact of your campaigns.

Would you recommend measuring incrementality on your own or with a third-party company—or somewhere in-between?

I encourage brands to develop their own solutions. After all, it is YOUR money, and the measurement tooling you run must be up to your standards.

While it is possible to measure incrementality on your own, and Foundry has helped brands get there, it is often beneficial to work with a third-party company that specializes in incrementality measurement. Here’s why:

• Expertise and tools: Third-party companies have expertise in designing and executing incrementality tests. They can access advanced tools and methodologies that provide more accurate and reliable results. Their experience in analyzing large datasets and understanding statistical models ensures that measurements are conducted effectively.
• Unbiased and objective insights: Third-party companies provide unbiased and objective insights into the incremental impact of your campaigns. They are not influenced by internal biases or vested interests, allowing for a more impartial evaluation of your marketing efforts.
• Scalability and efficiency: Third-party companies have established processes and infrastructure in place to handle large-scale incrementality measurement. They can efficiently analyze and interpret the data, providing timely and actionable insights.

That said, it is important to find the right balance between in-house measurement capabilities and third-party expertise. Some advertisers may build internal measurement capabilities while leveraging third-party support for more complex analyses or to validate their findings.

Since leaders like to look at annualized numbers, but we don’t know decay, how can you scale the incremental numbers over a year? Any best practices?

• Calibrate attribution models: Data from incrementality tests is used to refine and calibrate attribution models. By understanding the incremental impact of different channels and tactics, you can adjust the attribution weights assigned to each touchpoint in the customer journey. This ensures that the attribution model accurately reflects the true contribution of each channel and tactic.
• Consider seasonality: Take into account any seasonal variations in your industry or market. Adjust the scaling of incremental numbers based on historical patterns during specific periods.
• Allocate budget: Use insights from incrementality tests to allocate budget toward channels and tactics that drive the most efficient incremental cost per acquisition (iCPA). By identifying channels and tactics that generate the highest lift and incremental conversions, you can prioritize budget allocation accordingly. This helps optimize marketing spend and maximize return on investment.
• Re-measure optimizations: Continuously re-measure the impact of optimizations based on the results of incrementality tests. By implementing changes to your campaigns, such as adjusting targeting parameters, creative elements, or bidding strategies, you can evaluate how these optimizations contribute to greater incremental contribution. This iterative process enables you to refine your strategies and make data-driven decisions to drive incremental growth.

Generally speaking, do you think online brands with smaller brand awareness can put more value or trust in incremental lift tests in some ways? Or do you find that even smaller brands can run into the same issues?

• Smaller brands may be in a better spot vs. big brands in that the smaller brands have less brand equity. So theoretically, ad spend should provide more of a pop because ad media is the only way some folks will come to know a new brand.
• Incremental lift tests can help smaller brands identify the specific channels, tactics, or campaigns that are driving incremental results and optimize their marketing strategies accordingly. By measuring the lift in conversions or actions compared to a control group, smaller brands can gain insights into the true impact of their advertising efforts and make data-driven decisions to allocate resources effectively.
• However, it is important to note that smaller brands may still encounter challenges similar to larger brands when conducting incremental lift tests, such as ensuring proper test design, data quality, and statistical significance (i.e., amount of data). It is crucial for all brands, regardless of size, to plan and execute their incrementality tests carefully to obtain reliable and actionable insights.

Do holdout biases exist with geo-based holdouts, or is this exclusive to audience-based splits?

This is likely so, but geo holdouts can be a useful tool. Assuming that two markets move in concert, if one is treated with media and the other goes dark, the marketing effect lift may be understood as the performance of the two markets in direct comparison. Note that this usually takes an inordinately large amount of spend in the target market and having to go dark in the control market, which isn’t how you would actually run the campaign, so the results may not actually reflect reality.

While geo-based holdouts can be a useful tool for comparing the performance of different markets, it is important to consider factors such as market dynamics and the significant amount of spend needed in the target market.

Conducting holdout tests in a way that accurately reflects real-world campaign execution can be challenging. Going dark in the control market may not truly replicate how the campaign would actually be run, which can introduce biases and affect the validity of the results obtained from the holdout test.

What are your thoughts on a cross-screen campaign? Using an always-on approach and multi-touch attribution, we would want to tell the power of our multiple products.

• A cross-screen campaign with an always-on approach and multi-touch attribution can be a highly effective strategy to showcase the power of multiple products. By maintaining a consistent presence across various screens, you can engage with your target audience at different stages of their customer journey. This allows for a more holistic and integrated marketing strategy—maximizing the exposure and impact of your multiple products.
• Multi-touch attribution enables you to understand the contribution of each touchpoint in the customer journey and measure the incremental value generated by each product. It provides insights into how different screens and touchpoints work together to drive conversions or actions, allowing you to optimize your campaign and allocate resources effectively. To execute a successful cross-screen campaign, it is important to have a robust measurement framework in place, including proper tracking, attribution mechanisms, and advanced analytics tools. Continuous monitoring and analysis of campaign performance will help ensure its effectiveness.

Got more questions on measuring incrementality and lift?

Are you looking for expert consultation on your incrementality testing strategy? Need help measuring the lift of your campaigns? Get in touch with the Kochava Foundry team for an expert consultation.

Answers to your questions from the Kochava Foundry webinar

Grant Simmons, VP of Kochava Foundry, recently hosted the webinar “All Things iOS,” where he unpacked topics including Apple’s SKAdNetwork (SKAN), the AppTrackingTransparency (ATT) framework, Apple Search Ads, and more. The Foundry team gathered some of the most engaging questions that came up during the session to answer in more detail.

Check out the full webinar on demand here.

What are Apple’s plans with privacy manifests? What do you make of the list of SDKs they published? Will ad monetization SDKs be impacted?

Apple’s plans with privacy manifests provide more transparency to users about the data collected by apps and how it is used. Privacy manifest also empowers app developers by providing them with new control over whether third-party software development kits (SDKs) can transmit signal off device when user consent isn’t granted through Apple’s ATT framework. This level of control means that developers can safely integrate third-party SDKs, potentially avoiding the dreaded “contact your SDK provider” to change the default behavior of the SDK.

The list of software development kits (SDKs) published by Apple has a mix of well-known SDKs and obscure outliers. It appeared to be an addendum to a previous developer-focused communication. While the spirit seems to be attenuated to some SDKs that collect personally identifiable information (PII), it is important to note that the list hasn’t evolved much since it was first released. However, it is possible that more updates and additions will be made in the future. As for the impact on ad monetization SDKs, it is crucial for brands to review the list and ensure compliance with Apple’s guidelines to avoid any potential impact on ad monetization strategies.

I would love to hear about brands that are smaller advertisers and may not meet the thresholds of SKAN—what are alt models to attribution?

For smaller advertisers who may not meet the privacy or crowd anonymity thresholds of SKAN, alternative models to attribution can be utilized to understand campaign performance.

The first recommendation is to maximize owned media opportunities by hydrating special, usually hashed, first-party identifiers to the extent possible. This allows for better tracking and attribution within advertisers’ own media channels.

Another suggestion is to use network and publisher cost as an input and conversions as an output, implementing an MMM (Marketing Mix Modeling) or contribution model. This approach can provide valuable insights into the effectiveness of different marketing channels without reliance on deterministic, last-touch attribution. Exploring solutions like our AIM (Always-On Incremental Measurement) product can offer further guidance and support for smaller advertisers seeking to future-proof their attribution efforts.

How are you thinking about over-attribution related to SKOverlay elevating impressions to a Fidelity Type 1 in the SKAN waterfall (30-day attribution window)? Tangentially, how should brands navigate the loss of user intent signal with networks individually defining/justifying passing impressions as clicks on the mobile measurement partner (MMP) side with SKOverlay signaling and Engaged View type “click” signals? Ultimately, this becomes a measurement problem, but how can brands partner with their MMPs to address?

Over-attribution related to SKOverlay, which elevates engaged view impressions to a Fidelity Type 1 in the SKAN attribution waterfall, can be a challenge for brands. The passing of impressions as clicks on the MMP side, especially with SKOverlay signaling and Engaged View type “click” signals, can lead to inaccurate measurement and attribution. It is important for brands to recognize the misalignment that can occur, particularly with in-app ad units that generate multiple “tap-like” activities without significant user interaction.

To navigate this issue, brands should work closely with their MMP to establish clear guidelines and processes for differentiating between impressions and true user engagement. Standardizing reporting across offers and DSPs can also help address this measurement problem, allowing for better normalization and accuracy in attribution.

With Apple building their own DSP and possibly an SSP as well, what is the future for third-party ad platform access to iOS?

The future for third-party ad platform access to iOS may become more restricted with Apple building their own DSP and possibly an SSP. Apple’s focus on privacy and control over user data may lead to tighter restrictions on data sharing and targeting capabilities.

It is conceivable that Apple may push for a more direct integration with publishers, effectively making them integrate directly with Apple and requiring all internet traffic to go through the iOS operating system. It is important for brands to adapt to these changes and explore alternative advertising strategies, such as leveraging Apple Search Ads and other first-party data sources, to ensure continued access to iOS users.

What do you think Apple will release at WWDC 2024? And what do you think about broader private relay implementation?

While the specific releases at Apple’s Worldwide Developer Conference (WWDC) 2024 are uncertain, it is expected that Apple will continue to focus on privacy and user control. They may introduce updates to their privacy features, additional components to privacy manifest, data granularity improvements to the SKAN framework, and new tools for developers and advertisers.

Apple is aware that there are still many unsolved challenges in the advertising space, and they have already implemented visibility around Apple Search Ads that Google cannot match. As for the broader private relay implementation, it is an interesting development aimed at enhancing user privacy by masking IP addresses. However, its broader impact and adoption are yet to be seen, and it will be important to monitor how this aligns with industry practices and regulations.

Got more questions on iOS?

Need help with your ATT prompting strategy, SKAN or Apple Search Ads, or any other iOS marketing topic? Get in touch with the Kochava Foundry team for an expert consultation.