Ad tech firm AppNexus on Tuesday said it is making its streaming data feature available on a limited basis. The platform enables real-time decision making, and the streaming data feature will be available to both buyers and sellers. The feature will enable AppNexus customers to access campaign data in real-time, a departure from batch data processing, which has thus far been the industry norm. AppNexus is testing the feature with select clients which it declined to name.
The company claims the feature will be especially helpful to programmable media buys, since it has a high degree of consumer customization.
"We believe [the streaming data feature] establishes a new table stakes for all ad tech platforms. Those that cannot offer real-time data streaming won't be able to enable the
kind of personalized and nimble targeting that is required for advertising across the programmable internet," Andrew Eifler, VP of product management for AppNexus’ Advertiser Technology Group,
told Real-Time Daily via email.
Contrasting streaming data to batch data, Eifler stated: "Batch data processing was adequate in the age of the programmatic internet but it doesn't work for the programmable internet, which is highly customized and dynamic. Streaming data allows advertisers to react to campaign performance moment-by-moment, rather than hour-by-hour.”
Batch data processing requires that marketers and publishers wait a period of hours to see campaign results. With streaming data, results on 70% of served impressions arrive in less than 10 seconds, while results on 90% of served impressions arrive within 30 seconds. AppNexus claims that by using its streaming data feature, marketers will be able to determine whether their budgets are reaching the right sites or targeting the right audience segments. They’ll be able to redirect ad spend to other platforms, devices and formats in real-time. By gathering continuous, event-based data, they can derive statistical correlations and build "closed feedback loops" that recalibrate campaigns in the moment.
"In the same way that the internet itself has become highly customized, digital advertising is increasingly powered by real-time algorithms that engage consumers at the
individual level, rather than by broad segment. Batch data processing is only so useful in this process. Marketers need a real-time data feed to continually recalibrate their campaigns," Eifler added.