As programmatic audio ad sales grow, one of the downsides to the automated buying has been a bit messy in terms of how spots are categorized. Now the podcast platform Spreaker is leaning into a technology solution to help with the cleanup of ads that are being programmatically inserted into podcast content. At the same time, the move should help give brands more comfort the placements of their ads are meeting brand safety parameters.
Spreaker is relying on what is has dubbed as M.A.G.D.A., or Machine Augmented Guard for Dynamic Advertising. M.A.G.D.A technology works by transcribing programmatic ads that go through the Spreaker ad marketplace in real-time. Spreaker says it has created machine-learning models to auto-categorize ads. In addition, the technology can also detect miscategorized ads. For example, if a political ad is miscategorized as fast food, M.A.G.D.A. will flag the error so it can be fixed.
Ad categories remain one of the key guardrails in programmatic advertising since they help direct where ads can and cannot be placed. The possibility to block specific Interactive Advertising Bureau categories from playing on a podcast exists already. But Spreaker points that if an ad is uncategorized or miscategorized, it can be fixed, therefore making the need to block redundant. In the long run, the aim is to help grow the amount of money brands invest into audio advertising.
During the past several months, Spreaker has been quietly beta-testing M.A.G.D.A. in a real-world application with Audioboom. During the 2022 midterm elections, the two companies used the technology to keep tabs on political spots. Spreaker says the results have “been impressive” as misleading or miscategorized ads were identified and have been much more easily traced back to their source. During the test, it says fewer than ten miscategorized ads escaped the first pass of the process during the entire campaign season. Yet at the same time, the ten were also much more readily identified as having the wrong categorization.
"We've always worked to bring accuracy, brand safety, and value for podcasting to the fore. When clients and publishers can trust what they're getting, it improves the whole sales cycle and generates more revenue,” says Lance Paterson, VP of Research at Audioboom. “This new technology steps up our game ahead of other platforms, gives our partners more reassurance, and our listeners a much more enjoyable experience.”
Audioboom is just one company connected to the Spreaker programmatic ad marketplace. It says a number of other companies are using its technology, but no names are being released.
Spreaker also shares that the name of M.A.G.D.A. also has a fun origin story inside the company. While the name of the technology is an abbreviation for Machine Augmented Guard for Dynamic Advertising, it says there is also a Spreaker team member of the same name. Magna Freitas was the first person to work on ad quality control for the company. Today, Freitas leads the division dedicated to ad quality. Spreaker says it felt fitting to name the technology after her.
Comments