Briefing: What a 30-Second 'Play' Does to Your CPM

Last week two measurement efforts landed at once and the industry spent a few days talking about definitions. This week was the hangover, in the good sense: no new bombshell, just the AMP Accords and the IAB's v2.3 guidelines sitting in their comment windows while everyone quietly works out what the new rules cost them. That is where the real money decisions get made, so that is where the desk spent the week.

What a thirty-second play actually does to a CPM

Take the headline proposal seriously for a minute. If an exposure only counts once someone has listened to or watched thirty seconds, then a slice of what you buy today as reach stops counting tomorrow. Downloads have always included files fetched by apps and never played, plus the fast skippers who bail before the pre-roll finishes. None of that disappears from the world; it disappears from the numerator. A show selling a 25,000-download episode at a $25 CPM might be selling 18,000 or 20,000 qualified plays once the threshold applies, and the same dollar rate now describes a smaller, truer audience. Your cost per actually-heard exposure goes up on paper even though the show did not change.

The desk's read: this is good for buyers and good for honest sellers, and awkward for everyone in between. A consumption metric rewards shows people finish and punishes inventory that was padding its count with autoplay and abandoned downloads. It pairs naturally with the response data, where the money is ultimately settled. Magellan's most recent benchmark put the average podcast purchase-conversion rate at 5.22 percent; a play-based denominator makes a number like that mean more, because you are dividing outcomes by people who were actually there.

Video wants into the feed

The other half of the AMP definition, that a podcast is anything that works with your eyes closed, quietly invites video into the same category the open ecosystem has spent twenty years building in RSS. The response taking shape is video carried inside the feed itself rather than surrendered to YouTube wholesale, and that thread ran through the week's trade coverage, including the Podnews weekly review. It matters commercially for one reason this desk keeps coming back to: a baked-in host read travels with the file. If the video lives in the same feed as the audio, the endorsement you bought reaches the viewer and the listener as one buy. If the video lives only on a platform, you are back to negotiating two relationships and hoping the read shows up in both.

A fair question about who gets paid

The quieter debate this week was about creators, not metrics: whether the people whose shows generate all this measurable attention are compensated in line with what the attention is now worth. It is the right argument to be having as the numbers get better. Better measurement cuts both ways. The same data that lets a buyer stop paying for never-heard downloads lets a podcaster prove that their finished, trusted, screen-free audience is underpriced. Standards do not settle that on their own; they just give both sides better evidence to negotiate with.

What the desk would do this week

Buyers: pull one live flight and recompute its reach as if only thirty-second-plus plays counted. You will not have perfect data, but even a rough haircut tells you which of your shows are selling audience and which are selling downloads. Price accordingly at the next renewal. And if you have a comment to make on the IAB draft, the window closes August 19; after that you are living with a ruler other people wrote. Sellers: instrument completion now and start reporting it before a buyer asks, and if you have a video version of your show, get it into your feed rather than orphaning it on a platform. When the play becomes the currency, the shows that can prove people stay are the ones that keep their rate.