A Luddite's Guide to Podcast Ad Vetting: Non-AI Slop to Get to The Top.

My Ad Buying Framework: A Simple Guide to De-risk Your Buys and Increase Your Confidence

By the end of 2025 I’ll have five years in the chair as a podcast media buyer. That’s tens of millions in DTC brand budgets managed and teammates I’ve trained now running million dollar campaigns on their own.

When I started with $5K across two shows, I got lucky. A couple of early fluke wins launched my career. Then came the spaghetti throwing. Some of it stuck. A lot of it didn’t.

Fast-forward to today. I lead the media buying team at ADOPTER where we’ve deployed over $100M in profitable, measurable dollars for our clients. It’s a bit of a luddite process, but hey, it’s honest work.

Through testing, training, and too much time in Google Sheets and Podscribe dashboards, I’ve boiled podcast buying down to a five-level vetting framework. It de-risks campaigns, provides a valuation for every buy, and builds buyer confidence at scale.

I did a daily breakdown on LinkedIn of each step a few weeks ago, but this is the full 1-stop shop for it, plus more nuanced and detailed stories.

If you are a buyer, this sharpens you. If you are a seller, this may explain why I skip your show like the plague.

Level 1: Brand Renewal History & Testing Cadence

Truth: media buyers are creatures of habit. Every agency has its own test formula, but the shape is the same.

I run three-episode tests, 30 to 45 days apart, always with a 30-day cancel clause.

• Episode 1: If it tanks, we pull the plug. It has to be bad. Real bad.

• Episode 2: If there is more traction, we continue. If not, we cancel.

• Episode 3: If breakeven or better, we ride the long tail. If stellar, we renew quickly.

For renewals:

• If strong, book 1–2x per month through year end.

• If borderline, re-test with 3 more episodes.

With this context, you can start to evaluate how brands are buying a specific show and deconstruct performance.

Example: The Moth.

• Quince ran six months, paused, then returned. That signals a durable, higher-income female audience.

• Lumen performs here. NetSuite, Shopify, HelloFresh, and Nutrafol struggle. That might suggest a suburban, mid-income base that is less startup-oriented.

Any brand spending more than 3 months is worth watching. More than 6 months is a strong signal. Annual deals are gold, though assume they negotiated a lower rate than you will.

Pro tip: ignore giant spenders like BetterHelp or Toyota. These are akin to brand awareness plays or or playing a different game on pricing than you will be able to until volume has been reached. They are poor ROI indicators.

When I train buyers, I spend a lot of time here. We dive into brand-by-brand renewal history, look at demo alignment, and even listen to the ads. This is critical work. Without it, you are gambling.

Years ago when buying for Füm, I started to follow after Sean Frank’s “Ridge Wallet” as well as “Sheath Underwear” as guide brands, because we had found during that season, men aged 40+ were low cost, higher likelihood to smoke, etc. And I knew Ridge and Sheath were selling well to those demos—and Sheath/Ridge were very ROI driven so I could trust their renewals. That led me into millions of dollars spent on comedy podcasts I wouldn’t have otherwise tested.

Level 2: Demographics and Psychographics

Some would argue audience fit is where you should start—I disagree. More often than not you’ll make the majority of your buying decisions on brand renewal history.

First I weed out the trash through brand renewal history then based on the survivors we can isolate based on audience demos and then allocate to our clients. When we look at demos there’s three ways we do it:

Pod Direct Demos (from the creator/network and their platform or survey based analytics).

Pod direct demos include:

• Age

• Gender

• Geo location

• Sometimes household income or other surveyed responses

Transunion data (via Podscribe)

Transunion data is IP matched data from credit applications to the audiences. It’s occasionally helpful, but not a huge priority. More affirmative and a better tool for look back attribution than for vetting IMO.

With this we can get more household data/income data.

Inferred data (content, brand renewals, and gut).

Honestly, this is where I focus most. Just use your head, dummy. If the show is a show about menopause, hosted by a medical doctor with expertise there do you think the audience is comprised of 20-something men? I assume not.

I ran a podcast for years called, “The Brewview.” It was a podcast centered around getting to know the human stories behind kendama’s most influential players. I found that conversation was better over coffee, so coffee became the anchor point of the show. A coffee sponsor made perfect sense.

My audience may have been 20-30 men mostly in America, but it would have flopped for a brand like Folgers. This audience was outcasts, mostly people looking to be unique and expressive. Onyx Coffee Lab, the sponsor I took on was perfect—they believed that we should “Never Settle for Good Enough.” Something Kendama players take seriously in their craft and I do as a human (and in my brewing process)—it' was extremely value aligning. They naturally would never have picked my show based on my demos, but the psychographics of the audience was perfect (they got a 10x ROAS on my show monthly).

Please don’t overcomplicate audience demo data. Use your gut, it’s probably pretty good.

Level 3: Engagement Ratios

This sounds secretive, and perhaps it was but it’s so simple it feels weird keeping it private.

Years ago, Bryan Goldmark found that on average shows would receive approx. 5% of it’s episodic listeners as apple iTunes reviews. Inversely, you could take the iTunes reviews and multiply by twenty and project the episodic audience size.

All we did was compare the estimate to the network provided data to establish a ratio.

i.e. Adam’s podcast has 500 reviews. We project 10,000 downloads per episode. Adam’s data shows his show actually only doing 5,000 downloads. That means Adam’s audience leaves reviews at a rate higher than average, or quantifiably a 2.0 on the “Goldmark Ratio".” They must really like his show, therefore are very engaged—why wouldn’t they be, I’m pretty great.

This is the principle of the formula.

Take Serial or Adam Carolla’s show for example. They’ve got tens of thousands of reviews from the many many years of podcasting which may not be indicative of today’s engagement or audience size. I wouldn’t use this formula for very old shows or very new shows. It’s too nuanced.

But for shows with limited ad history, it’s a nice additional tool to use to see if there’s value in the show or a potential inflation of downloads (I’ve caught some “paid for audience” shows using this).

Use this carefully, and don’t over index on it. Spotify taking a larger share of listenership also throws a wrench at this and I haven’t developed a new ratio to incorporate Spotify reviews—if you have thoughts on how to incorporate Spotify, I’m all ears.

P.S. after beginning to write this months ago, Podscribe launched a feature on their app for this exact scoring. They give a little more generosity than I to what they consider good, but it’s still incredibly helpful for quick scoring.

Level 4: Insertion Types, Ad Quality, & Placements

Alright, to be honest this is probably the area I overlooked the most early in my buying career and now maybe make too big a deal out of.

Insertion types matter. Live vs. Baked vs. Faked vs. Full Catalog etc. matter.

Each has a value inherit to it. I think in descending order on an impression vs. impression basis. The gap is much smaller today than it was when I first started.

But know what you’re buying and know that there’s differences in value.

Ad quality is mostly subjective—but I think you can feel what resonates.

Listen to the show’s other ads. Do they feel genuine? Would you consider buying the product if you were the right demo and fan of their show? Are they selling or just reading a script?

And lastly placements.

I’m gonna stop here and reiterate my consistent loathing of ad practices done by one of the major entities in podcasting… you know who… and other networks who choose to invade the listener’s experience with 20 ads in an episode. You foul monsters. Respect the audience—the only winners from that high of ad load or aggressive ad stacking is you, the network, profiting off of negligence and burning out shows. You should be ashamed if you claim to love podcasts and do this to them.

We are turning podcasting into a worse industry, please stop screwing up this place we built so well.

Ad practices like this, and even this new news about an AI content puppy mill coming get under my skin—maybe I am a luddite, but these are the practices where the deadly sin of Greed lives in. They seem profitable, but they kill value fast and leave nothing positive in return.

Though I wouldn’t say I’m very religious, I do believe in the value of stewardship I was raised with in my Christian home. We should leave things better than we found them. We should not over reap the soil in which provides us fruit. Not everything needs to be milked dry.

Be better. Buyers can speak with their dollars and their voices.

I digress.

Level 5: Negotiation Tactics

Lastly, you pool it all together.

PSA: If you’ve worked with independent podcasters, they’re often trained or pushed to ask brands’ for their “budgets.” This is a dumb conversation for performance oriented brands. Ignore it until you have details on potential deliverables.

Vetting shows gives you confidence. Negotiation is where you cash it in. This is where we stop gambling on rate cards and start stacking the odds in our favour.

1. CPM is the Anchor, Flat Rates Are the Output

Networks love quoting flat rates, but those only matter once we know the CPM logic behind them.

CPM is the constant.

Flat rates flex with impressions.

CPM is the number that I value the audience against. Is it s $25 audience or a $40 audience? Only vetting will tell you, not a sales person.

2. Terms That Protect Performance

Deals aren’t just about price, they’re about conditions. Smart terms de-risk your spend:

30-day cancels or better → always. 14 is reasonable, anythign less is probably offensive and predatory.

Placement clarity → mid-roll 1 or equal rotation.

Reporting → impression logs, placement details.

Makegoods → terms to get one and what you’ll get. Impression guarantees, copy mess ups etc. You should get same length/placement within 30 days.

These guardrails let you scale without being handcuffed when things go awry.

3. Test → Confirm → Scale

Think in phases:

Test: 3 episodes, CPM anchored to risk.

Confirm: if decent or good 3–6 more episodes, same CPM unless performance forces a change.

Scale: Lock in cadence (monthly, or 20% share of voice), maintain cancel terms as a safety and keep testing.

If you maintain a 30-40% hit rate on shows you’ll have a scaling campaign. under 30% strategy is probably off or podcasting sucks for you.

4. Trade, Don’t Concede

Every ask should have a give.

“We’ll stretch CPM by $2 if you guarantee mid-roll 1.”

“We’ll commit to more episodes if you cap ad load.”

Never give something away without getting something back.

5. Card Counting, Not Gambling

Each vetting step adds a card to your hand—renewals, demos, engagement ratios, ad quality, placement. Negotiation is how you stack those cards into an advantage.

For example, a show like You Should Know (via QCODE) for Ridge Wallet.

Renewals: strong (Harry’s, Füm, Hims).

Demographics: young male skew → good, but not a premium niche audience like “Dr. Tyna”

Engagement: healthy engagement ratios → invested audience = confidence.

Placement: solid reads, but younger demo caps CPM.

I’d price it below $25 CPM. That’s not gambling—it’s an informed, de-risked test. All fronts look good, but the audience demos are male and young, and likely lower income. Probably great for Ridge at a $20-25 CPM or better.

The Point:

Don’t haggle random rates, anchor price through your vetting process.

That’s how you scale campaigns with confidence.

Good luck, and stay caffeinated.