New NumberEight research finds that demographic data advertisers trust may not be telling them who their audience actually is, with IP and household-level approaches failing to capture meaningful differences between audiences.

London, UK, September 23, 2026 – As media consumption moves beyond traditional channels into mobile, gaming, podcasting and CTV, advertisers need to understand audiences across increasingly fragmented environments. Yet the demographic data that underpins much of media planning and targeting still relies on methods built around persistent identifiers and household-level signals. New research suggests these established approaches may not accurately reflect the audiences actually consuming the content.

New research from NumberEight, a leader in AI-native audience intelligence, compares traditional approaches, including self-reported data, IP-based matching and third-party identifiers, with ID-less methods that infer demographic characteristics from contextual and behavioral signals in combination with anonymized first-party data.

Drawing on existing research alongside real-world mobile and connected TV (CTV) analyses, the report examines the strengths and limitations of each approach, with particular focus on how demographic accuracy can be affected when IP addresses are used as household- or network-level match keys for person-level attributes such as age and gender. This is increasingly relevant as other persistent identifiers become more restricted and IP plays a larger role in identity resolution.

Different content, remarkably similar data

In mobile gaming, NumberEight compared gender predictions using its ID-less methods to an IP-based approach from a leading US data provider across games with distinctly different audiences. The IP-based model consistently returned gender distributions close to 50/50, regardless of the game.

For example, Fashion Battle was identified as 51% female by the IP-based model, compared with 82% female using NumberEight’s ID-less model. Nitro Nation, a game expected to skew male, was identified as 50% male using the IP-based model, compared with 74% male using the ID-less model.

A similar pattern emerged in CTV. NumberEight analyzed a third-party dataset covering 7,013 shows and more than 3 million unique users. Despite established differences in consumption habits between genders, household-level demographic data showed very little variation. Gender remained tightly clustered at approximately 53% to 55% female, while age distributions were also strikingly similar across content ranging from Animation and Comedy to News and Documentary.

The findings highlight the limitations of using household- or network-level match keys to infer individual demographic characteristics. An IP address can help connect activity to a household or network, but it cannot reliably determine which individual is consuming the content.

“Demographic data is one of the most widely used inputs in advertising, but we rarely stop to question how that data is actually generated or how well it reflects the person consuming the content,” said Emma Raz, Commercial Director at NumberEight. “What stood out in this analysis was how consistently some traditional approaches produced almost identical demographic profiles across content we know has very diverse audience profiles. The issue isn’t simply whether you can match an identifier. It’s whether the data attached to that identifier actually tells you something meaningful about the audience.”

Rethinking demographic intelligence

The research also examines the wider limitations of self-reported data and third-party identifiers, from inaccurate or false information to identity graphs becoming fragmented as cookies disappear, mobile IDs become unavailable and users move between devices.

By contrast, NumberEight’s ID-less approach uses anonymized first-party demographic data alongside contextual, behavioral and device-level signals to infer demographic characteristics without relying on PII, IP addresses or device IDs in the processing. Across the mobile and CTV analysis, this approach showed greater variation between audiences and more closely reflected expected differences in audience composition.

“The shift away from persistent identifiers creates an opportunity to rethink how we understand audiences,” said Abhishek Sen, CEO and Co-founder of NumberEight. “Sophisticated AI-led approaches that combine a variety of data sources from various channels enable us to create holistic predictive models that cover demographic predictions that are omnichannel at their core. Instead of treating identity as a proxy for understanding, we can focus on whether the signals actually tell us something meaningful about the audience.”

You can download the full Demographic Data in Advertising: Traditional vs. ID-less Methods report here.