Frequency capping has long relied on a static “set it and forget it” mindset: once every 24 hours, three times every seven days and so on. This passive approach to avoiding consumer ad overexposure relies on the presence of ad identifiers, which continue to deteriorate.
New approaches like performance-based frequency use advanced modeling to determine optimal ad exposure timing and frequency for consumers, considering their unique journey across channels. Importantly, this approach can reduce average user ad exposure while simultaneously improving campaign performance.
By delivering ads at the most impactful moments, consumers avoid seeing the same ad repeatedly across channels, which yields better results for advertisers.
Let’s break down the science behind the solution.
Frequency management’s long-overdue makeover
Marketers are well acquainted with the shifting landscape of addressability amid rising privacy expectations. But when it comes to frequency management, these changes highlight just how limited traditional exposure strategies have become:
- Incomplete measurement. Simplistic frequency caps fail to capture a unified view across diverse media channels.
- Overemphasis on delivery metrics. This forces advertisers to make artificial trade-offs between budget delivery and frequency control.
- Fragmented campaign management. This results from limited visibility and siloed controls within DSP platforms.
Collectively, these issues create a disjointed approach to frequency management that is out of date with the complex, multi-device journey of today’s consumers.
A better way: Performance-based frequency
Performance-based frequency is about achieving better performance and outcomes for everyone: consumers, advertisers and publishers. It does this through what we call total exposure modeling, a new way to model and execute frequency management that unifies three critical elements:
- Holistic view of ad exposure: New frequency models capture unique consumer journeys by understanding their ad exposures across all channels: traditional digital ads, streaming TV, linear television and more. For the first time, they integrate linear TV impressions into digital frequency management through advanced modeling.
- Probable exposures without ad IDs: Traditional models consider known exposures we can directly measure. Total exposure modeling also measures probable exposures where traditional ad identifiers aren’t available, predicted future exposures based on aggregated behavior patterns and the cumulative impact across all channels.
- Performance-based frequency bidding: By using frequency as a fundamental input for bidding decisions, advertisers can move beyond rigid caps toward optimal exposures, enabling context-aware and dynamic bidding based on frequency intelligence and real-time user behavior. The bidding system should determine optimal frequency by considering both past ad exposures and future impression opportunities.
How to make it work
Signals are the engine of modern performance-based frequency management, enabling advertisers to make intelligent decisions. As the industry shifts away from individual ad IDs, these signals become increasingly crucial for understanding and optimizing ad exposure.
Here’s how to effectively use exposure modeling for frequency management:
- Reach your audience based on available signals: Use a cohort-based modeling approach that makes intelligent frequency decisions without relying on ad identifiers, using cross-channel and full-funnel signals as inputs.
- Make informed predictive frequency decisions: Find the ideal exposure level for each campaign by looking beyond immediate metrics like clicks and conversions to understand how frequency impacts long-term effectiveness. Total exposure modeling achieves this by balancing immediate performance with optimal frequency to find the sweet spot between effective reach and ad fatigue.
- Validate your model: Panels can reveal actual behavior across all traffic types, including sessions without ad IDs. This means modeled frequency decisions can be validated against real-world data. One way we have chosen to do this is by building a panel of opted-in Amazon users, who populate the Amazon Shopper Pane.
Does the model work? A short case study
Recent internal testing demonstrates that campaigns do achieve optimal performance at specific frequency thresholds. At Amazon Ads, one test compared a control group with no target frequency applied to the algorithm against groups with various optimal target frequency settings.
The control group had a median CPC (cost per click) of $0.775, while the treatment (optimal) group achieved $0.656, representing a 15% decrease in CPC savings for the advertiser, resulting in more working media dollars for marketers.
Crucially, the data revealed optimal performance at two exposures per user efficiency declined at both lower and higher frequencies. By optimizing the timing and delivery of these exposures, we achieved greater impact with fewer impressions. The data demonstrated that smarter frequency management outperforms simply increasing ad exposure.
Amazon Ads advertisers are seeing an average of $52,000 reinvested by using optimal frequency caps at the order level and 17.6M impressions saved annually.
This proven frequency optimization not only helps reduce ad fatigue among viewers but also allows advertisers to distribute their budgets more strategically across their desired audiences.
The keys to successful performance-based frequency management
Innovations like total exposure modeling are an essential tool in an over-saturated ad environment that lacks ad IDs. However, they require access to a DSP that provides cross-channel signals that enable the holistic view the model relies on. Understanding the customer journey across a variety of touch points, from ecommerce interactions to streaming and more, at scale, is critical.
The old ways of using frequency caps are disappearing, never to return. For advertisers, publishers and, especially, consumers, that’s a good thing.
For more articles featuring Pieter de Zwart, click here.
