By: Tylynn Pettrey, SVP, Analytics & AI at Chalice AI
For years, digital advertising has been guided by a simple principle: optimize what can be measured. That approach helped scale the industry, but it also shaped the behavior of the systems behind it. Over time, algorithms became highly effective at identifying and prioritizing signals that were easy to track, even when those signals had only a loose connection to real business outcomes.
That distinction matters more now than ever.
Metrics like clicks, last-touch attribution, and viewability created a shared language for performance, but they also introduced a kind of shorthand. These signals offered clarity and consistency, which made them useful, but they also narrowed the definition of success. As a result, optimization often centered on moments that were easiest to capture rather than those that were most influential in driving growth.
When algorithms are trained this way, they begin to favor immediacy over impact. Media is directed toward users who are already close to taking action, and performance becomes closely tied to timing rather than influence. The system, in effect, becomes very good at recognizing intent without necessarily creating it.
A shift is underway as more advertisers begin to rethink what their models are actually optimizing toward. Instead of relying solely on proxy metrics, there is a growing emphasis on outcomes that reflect genuine business impact, such as incremental sales lift, household penetration, or the acquisition of new customers. These measures are more difficult to capture, but they provide a far clearer picture of whether advertising is doing its job.
When optimization aligns with these kinds of outcomes, patterns begin to change. Environments that support deeper engagement tend to perform more consistently. Context plays a larger role in shaping results. The quality of the surrounding content, the structure of the page, and the overall user experience all become relevant signals in predicting performance.
This is where a more granular understanding of media begins to matter. A single domain can contain a wide range of experiences, each with its own level of relevance and value. Treating all impressions within that domain as interchangeable overlooks important differences in how consumers interact with content. Page-level intelligence offers a way to capture those distinctions, allowing models to evaluate placements based on the specific context in which they appear.
With greater precision comes a more rational allocation of spend. Inventory associated with strong editorial environments and thoughtful design may carry a higher initial cost, but it often delivers more reliable results when measured against meaningful outcomes. Over time, this creates a clearer link between quality and performance, reinforcing the value of well-crafted content.
For advertisers, this evolution changes how media investment is approached. Budgets are no longer optimized purely for efficiency, but for effectiveness over time. Measurement becomes more closely tied to independent signals that reflect actual consumer behavior. At the same time, flexibility remains important, as not every advertiser has access to the same depth of first-party data. Models must be able to adapt to a range of inputs, from retail sales data to panel-based insights.
The underlying technology has advanced quickly, but the more important shift is conceptual. Algorithms will continue to do exactly what they are trained to do. The real question is whether the industry is asking them to optimize toward the right things.
As that definition evolves, so does the market itself. Systems that prioritize meaningful outcomes tend to elevate higher-quality environments and more relevant connections with consumers. In that sense, the future of advertising will be shaped less by how efficiently it can scale, and more by how effectively it can align incentives with impact.
SVP of Analytics & AI at Chalice Custom Algorithms. I love to create innovative machine learning algorithms to tackle the needs of the ever-evolving programmatic advertising industry.
