The channel most B2B companies cancel too early.
Podcast attribution undercounts. If you judge a podcast buy on last-click, you'll kill a working channel and be able to show your work.
Does podcast advertising work for B2B? It works and it under-reports, which is why it gets cancelled. Host-read mid-rolls on niche shows whose audience matches your ICP perform well on pipeline and poorly on last-click. Measure with brand search lift and a self-reported attribution field, or don't buy it.
Where podcast advertising works, and where it doesn't.
Where it works
- Niche shows where the host has genuine authority with your buyer
- Host-read mid-roll placements, which are endorsements rather than ads
- Multi-episode flights that let frequency and familiarity build
Where it doesn't
- Large general tech shows bought on download count
- Single-spot tests
- Organizations that require last-click justification for every line item
Three expensive mistakes.
Buying downloads instead of audience.
A show with 15,000 downloads and a perfectly matched audience beats one with 500,000 and a general tech listenership, at a fraction of the cost. The big show sells reach; the small show sells your buyer.
Taking the produced spot instead of the host read.
A pre-recorded ad dropped into an episode gets skipped. A host saying the words gets listened to. Host-read mid-roll is the premium placement for a reason, and it's worth the premium.
Judging on a vanity URL.
Almost nobody types the URL. Promo codes undercount, pixel attribution through the networks catches a fraction, and last-click catches almost nothing. The honest reads are brand search lift during flights and a 'how did you hear about us' field.
What actually gets built.
Audience Match Over Download Count
Shows selected on listener fit and host authority, with download volume as a secondary filter rather than the first one.
Host-Read Mid-Roll
The endorsement placement, briefed with talking points rather than a script so it sounds like the host and not like a read.
Multi-Episode Flights
Three or more episodes per show across a four-to-eight week flight, because a single spot produces no measurable familiarity.
Lift-Based Measurement
Brand search volume tracked against flight windows, plus a self-reported attribution field at the form, treated as the primary read rather than a supplement.
The honest part.
If your organization can't accept a channel measured on lift and self-report, don't start. You'll fund it for a quarter, see nothing in the attribution report, and cancel something that was working.
And if you can only fund one spot, put the money into three spots on a smaller show.
What buyers ask first.
Related: Newsletters · Measurement · Growth Strategy
Find out what's actually broken.
Thirty minutes, direct with Jer. He shows up with findings, not discovery questions.
Not ready to talk? Run the Leak Report →