Podcast sponsorship buying, host-read and baked in

A download is not a listen, and the gap is worth money. Publishers measuring under IAB Tech Lab Podcast Measurement v2.1 report numbers that typically land fifteen to thirty per cent below what the same show reports uncertified. We buy against certified figures only, we pay for mid-roll rather than pre-roll, and we buy the show rather than the network's roster. What we buy and what we refuse: Host-read, baked into the audio file, sixty to ninety seconds, placed after the first content break. Baked-in reads keep working: a show with a back catalogue people actually work through will send traffic in month nine from an episode you paid for in month two, while a dynamically inserted ad is pulled the day the flight ends. We do not buy network run-of-roster packages. A network sells those to clear inventory nobody asked for, and a plan containing fourteen shows you cannot name is not a plan. The brief: We give the host talking points and a fact sheet, never a script. Three points, one of them a limitation of your product stated plainly. Hosts allowed to say what a tool is bad at get believed on what it is good at, and the read stops sounding like the ad break. We ask for the read to be anchored in a specific situation from the host's own working life, and we listen to the episode before approving the invoice to confirm the read happened and ran in the position we bought. Attribution, honestly: Promo codes and vanity URLs undercount badly, because most listeners hear the read while driving or training and search your brand name later. We use three sources together: pixel-based attribution on the publisher's feed through a provider such as Podscribe, a branded-search baseline taken four weeks before the flight, and a geographic holdout where the audience concentrates enough to allow one. The holdout is what settles arguments internally. If your target shows are small and internationally spread, we will tell you no holdout is available and that the buy will be judged on branded search and self-reported source instead. Where we work: Endurance, strength and sport-science shows, and developer shows. Endurance audiences buy equipment, coaching and software, and the host's own training is the context that makes a read land. Developer audiences punish marketing language faster than any audience we buy for; the read has to be technically correct or the episode comments say so within a day. We will not place a developer buy without an engineer from your side on the brief call, and we have turned down mandates over that condition. Not included: We do not produce a branded podcast, book you as a guest, or buy programmatic audio through Spotify Ad Network, Acast Marketplace and similar exchanges. That inventory is announcer-read, dynamically inserted and bought on CPM. It is a different job with different economics and a performance shop will run it for less than we charge. Who this is not for: Products with a purchase cycle shorter than a fortnight; this channel works on people who hear something and act three weeks later. Brands unwilling to let a host disagree with them on air. And anyone who needs a cost per acquisition agreed in advance. Nobody can honestly promise that on host-read audio, and a shop that does is selling you last-click credit for demand that would have arrived anyway.