The Pour Over: Q2 PPB

I spent the better half of a day deep diving into the Q2 Podscribe Performance Benchmark report. It’s probably my favorite piece of literature in our industry on a quarterly basis.

It’s helped refine my thoughts, and has been a good barometer for my feelings towards certain things—it’s put data to feelings.

In this send, I poured over (coffee pun) the PPB without any guidance or context and wrote along as I did to give unfiltered thoughts side by side.

Thank you Amelia, Pete, and Matt for your hard work on this as well as the broader Podscribe team.

Grab your mug and enjoy!

Methodology Shifts

The first issues of the PPB focused specifically on some of the seasoned advertisers. Brands that had run at least 5 million impressions in the quarter. Now they’ve opted to include advertisers with a minimum of 1 million impressions.

This has grown their sample size from 171 brands to 238 and total ad campaigns from 39,000 to 58,000.

My hunch is that this will slightly worsen some performance benchmarks but also help make numbers more realistic for new advertiser trying to grasp the state of the industry.

Additionally, one really good shift I think they made is in counter to my argument above. Instead of taking averages that may get skewed by the massive amount of campaigns from smaller advertisers, the benchmarks are based on a median number (the middlemost number in the data).

Keep this in mind as we dive deeper.

Conversion Rate Breakdowns:

Visitor rate

Q2 visitor rates are sitting at .33% meaning that if you advertised to 100,000 people, 333 would make it to your site. And if you hope to convert 3% of them on your site you’d have ~10 customers.

Say you paid a $25 CPM for that reach your cost would be $2,500. Your CPA would be $250.

If you’re a CPG product, this is not great unless your onsite conversion rates are great.

It’s helpful to remember that this is median data across a lot of brands. Most of the biggest spenders are high AOV products or B2B solutions like BetterHelp, SimpliSafe, Shopify, ExpressVPN etc.

These numbers aren’t far from accurate though. I have some clients higher (.77-1.5%) and some around .33% or lower but with on-site conversion rates over 10% + CPMs in the mid-teens.

It’s helpful to know, but should not control your expectations.

Install rate

.12% is the median install rate. Helpful for application-based products. I don’t really work with many, so I don’t have much to say here.

Purchase Rate

This quarter’s number: 0.021% 

This number is realistically showing full-funnel performance from impression to purchase.

If we take the visitor rate in mind, we can deduce the median conversion rate should be around:

If we expect our 100,000 impressions to drive 333, this number is saying we should expect 21 orders.

So from 21 orders, we can work backwards to get a 6.3% conversion rate.

With this, if we paid $2,500 for this ad we would land around an average $119.

Interesting!

Oh, that’s literally what the next slide says (I’m writing as I read through it… this is unfiltered).

Attributed Visitor Purchase Rate

So with our 6.3% on site conversion rate and an assumed $25 CPM, we should expect across a campaign a $119 CAC… Not.

It’s so different for every brand, but benchmarks are interesting.

Frequency

This is my favorite section as it says more about the industry as a whole than it does about individual campaigns.

This is one of those sections that gets me excited because it proves what we’ve been feeling and seeing.

High frequency does not necessarily mean more or better performance. Contrary to popular belief, you don’t need to hear an ad that many times to convert (so please, stop selling uncapped campaigns… it’s nonsense perpetuated by publishers to make more revenue off dumb brands).

The highest conversion rates/purchase rates come after just one ad and begin to taper after.

1. Firstly, this section does not separate episodic vs. impression-based. It’s just saying how many ads a unique user has heard/downloaded the ad.

2. The highest conversion rates are landing at the first ad with a .69% visitor rate and a .06% conversion rate (from impression). This drops nearly 40% at the second ad, another 40% to the third, and then tapers more slowly to the fourth and fifth.

3. This data feels a bit off to me, personally. I have some low-cost brands selling easy-to-buy products ($50-100) that convert pretty quickly–BUT, some products, $300+ take a few ads to convert better seeing peak conversions at the 2nd/3rd ad. Depending on the indexing of the survey and what the median is pulling more towards that could sway the data. Every product is different, but I’ve rarely seen incredible conversion on users after the 4th ad.

Episodic vs. Impression-Based Campaigns

I love the opening line in this section:

“Overall, episodic buys perform ~2x better than impression buys in terms of visitor rate”

I’d like to pat myself on the back for years of saying this alongside the army or performance buyers who’ve been measuring since day one.

But, the question is WHY?

In talking with Pete and the Podscribe team I think it comes to two things:

1. Frequency (as indicated by the above section)

2. Placement.

Placement with impression buys has much more volatility. Generally, your ads are at random placed in open markers across the catalog. Many are in later groupings of ads so it’s hard to predict where in an episode you appear. Many episodic spots get favored in first-position spots (thank you!).

Additionally, since there are still some iOS 16 devices, there’s inherently a leakage of performance (about 10-15%).

Lastly, and more just what I feel, I don’t think the people listening to back catalog episodes convert as well as people tuned into the content dropping now. They’re “lazy listeners” not fans.

They cherry-pick content but aren’t necessarily influenced as much by the show/host.

I think these factors will always dilute some of the performance on impression-based buys.

Note, this is anecdotal and just a hunch regarding my reasoning behind the data. Perhaps it’s for other reasons, but I believe this is a strong answer with some logic.

Personal Confession

Look, I know I haven’t been nice to DAI impression-based buys, and even DAI in general. It’s cost our clients a lot over the years. And now, after painstaking efforts,  it’s finally come closer to where it needs to be.

10% gaps in visitor rates after 1x impression in favor of episodic.

I am still skeptical.

Pixel data supports the advancement, but when I compare coupon data, UTM data, and Post Purchase Survey data, I often feel a wider gap.

It causes me to question some things. Why do impression-based buys not show up through traditional direct response tracking as much?

This is a hard question to answer.

Are people listening to older episodes not as in the moment engaged and are passively being influenced? Are they for some reason so wealthy they don’t care about savings?

Why?

Perhaps show notes matter more than we think. With impression-based buys, the code/URL aren’t generally visible in the show notes–so we’re relying on memory much more heavily.

This is really my only formal conclusion. Or there’s something deeper.

Host Read vs Produced-Read

I hate produced reads. I think they dilute our space and make it worse for nearly everyone. It feels like we’re putting up low-cost apartment buildings in beautiful beachside towns like Carlsbad, CA (heaven on earth). It makes everyone else’s property that much less desirable. It brings everything down–yet, I see the draw for both brands and creators.

Produced reads are what big brands want. Coca-Cola can’t afford to let Bill Burr rant on for 5 minutes about their brand. And yes, I’m in favor of big dollars coming into our industry. It helps make it a bigger space for everyone to thrive in–I just recognize the cost we are paying.

Anywho, that’s a rant for a different newsletter.

The host read still reigns king (and queen) for brands with nearly double the performance in visitor rate, and just shy of double for purchase rate.

There’s a healthy discussion to be had here. If you’re buying produced reads at 50% of the cost, you can assume near-similar results.

That’s possible.

Get a 1x/Month frequency, produced read, with a $10 CPM? Not a bad deal per se.

Just avoid the minefield of publishers who won’t frequency cap, jam 12 ads in an episode next to yours, and get these rates and you’ll be in a healthier camp.

What I think this section of the PPB does highlight that I’m still wrestling over is why the gap is over 60% on episodic in favor of host-reads and only 40% on impression buys.

I expected more uniformity.

The takeaway they want to leave you with is that 1st priority is insertion type then ad type.

Episodic insertion far out weighs impression-based purchase rate, whereas host-read vs. produced plays a small factor.

Placement

What about placement? I love that Podscribe captures this.

If you’re the first ad vs the last ad, how much does that matter? Turns out: a lot.

It’s clear there’s not a major difference in performance if you’re in the first couple of groups, but the moment you have your ads tucked in the 4th/5th groups (generally at 65% in or later performance wanes considerably.

Interestingly, there’s some evidence here to say pre-rolls are actually pretty solid only slightly falling short in visitor rates to the first mid grouping (group #2).

But say you’re in the first three groups, you should be happy… maybe.

Where in that group you’re placed matters the most.

God forbid you’re placement #4, at less than 50% of the performance as position #1. Who on earth is stacking 4 ads per group and over 14 ads per episode… I can think of a few. Ludicrous behavior.

Watch for severe ad stacking–it’s killing your performance.

Day of the Week

I’m going to skim past this. We can’t really control the day of the week except through programmatic. I think this more vanity than directional, but here’s the quick high-level:

If you can, hit them on hump day. Best visitor and conversion rates are on Wednesday and avoid them on Saturday–people are just out for brunch and vibin’ they ain't buying.

Ad Length

Yes, I suppose size matters.

Every 15 seconds you add to a read the better the visitor potential. But after 90 seconds purchase rate starts to diminish. I imagine these listeners are visiting to find the contact form where they can complain about how much of a podcast you took away from them.

“Why waste time say lot word when few word do trick?” - some guy who can’t hold chilli.

Genre Benchmarks

So much nuance here.

Apple’s classification for genres is not always the best descriptor–but we must progress forward.

Tech is king for visits and purchases–also CPM usually. These shows aren’t cheap. They’re not really for your granola bar brand. Big B2B/SaaS brands or tech products usually land here with high AOV/LTV options.

Business and Health and Fitness folow suit. Business is nearly identical in breakdown as tech.

Health and Fitness, however, I buy a lot of.

These creators are the best and worst. Expensive, nuanced, and complicated. They almost never stay on a script or within FDA guidelines. That said, they’ve got influence and the more fringe they are in their health views the better (for performance) often.

Religion is great as well–high influence like the Health and Fitness genre. Need I digress?

The major three you’ll see on most DTC brands: Comedy, True Crime, and News all rank fairly low. Why is that?

Little secret: they’re not selling at as high of CPMs…

News and Comedy especially.

True Crime has more misses than hits these days. Everyone seemingly has a degree in forensics and a story to tell. Not all are worth buying. Just be wary–though some are phenomenal.

ROI: Show Genre

I don’t think I like this section. It’s so complicated and I think it could be misleading. Just be smart about picking shows your brand aligns well with both the audience and host. Also check renewal rates, and negotiate well.

Ad Errors

Okay, say you’ve done everything in your power when you go to book your campaign well and it still doesn’t work… well maybe you need to recalibrate and you’re the problem OR there’s a myriad of issues in delivery that could occur from a bad read, a short read, or being stuffed in an ad stack.

Some matter more than others.

Short reads are near negligible on performance. Only 6.12 percent worse than normal whereas when you allow more than one ad to be in the same episode to the same brand (don’t make me say names) the performance tanks 59.18%.

Stacked ads are almost as bad (being the third or fourth ad in a stack) with a 40.82% worse performance.

Bad reads only factor up to 22% of performance. A great read will not turn a 0.25x ROAS campaign into a 2x ROAS campaign–stop trying to put a square piece in a round hole and take it out back behind the barn.

.The real culprits are on the publisher's side–please stop stacking ads and for the love of my vascular and mental health never run more than one ad for the same brand in an episode. 

Advertiser Industry Benchmarks

I’ll keep this light. Media is #1 for visitor rates. It’s not hard to get someone listening to a pod to check out another, check out a prime show, etc. Gaming is second. We’re addicts. I played World of Warcraft semi-professionally and it doesn’t take much to sell me on a video game–I promise.

CPG food and beverage is middle of the pack, and that’s most of what I buy next to CPG Health and Beauty. No wonder I come across as blunt and mildly pessimistic in a Scott Galloway-style fashion.

Don’t lean into this data too much. Set your own benchmarks.

iOS 17 Adoption Rates

Sept 18th is now a forever glass of wine night for the hard work put into fixing the nightmare DAI has been for podcasting.

We’re at near 90% adoption rate of iOS 17 which is in general why DAI impression-based buys have closed some of the gap to episodic campaigns.

For context prior to this update, individual devices could download 200+ ads and absolutely destroy a campaign before it started–as mentioned above a listener does not need high frequency to convert (don’t use this as a selling point to me).

For years, Glenn and I harped on DAI and how bad it was along with many in the industry on the performance buying side. We hadn’t figured out until last Spring why it wasn’t working. And with the help of Pete Birsinger, Brittany Clevenger, and Bryan Barletta we found the root of most of the issues.

Apple auto downloads. Dun dun dun…

Alas, after the Sept 18th introduction of iOS 17 the problem lessened.

We’re nearly 90% done with that nightmare, but it still impacts performance–just not as much.

Notably, this update CRUSHED some shows in our industry. It led to mass layoffs at companies like iHeart and the likes. The shows pumping big back catalog numbers milking ad dollars on downloads that weren’t real were hit the hardest. 

Revenue tanked and a slice of humble pie was dished.

I say this with a bit of petty pride and a slight childish demeanor of “I told you so.” However, I feel bad. Yet, the writing was on the wall.

Performance buyers lead the brand buyers. The brand dollars only exist because of performance buyers.

They may not be as plentiful, but they are the most important because they set the foundation for which the fluff our industry likes to sell rests upon.

So, I urge you, publishers and podcasters… do not ignore the pleas of performance buyers. They care more than you know about the success of our medium and the long term sustainability.

In Summary

Let me conclude:

1. Episodic is still #1

2. Host reads are still #1

3. If you sell impression based without frequency caps you are a predatory sales company

4. Ad placement matters a ton–stop running 4 spots back to back (preferably no more than two.

5. You don’t need a lot of ad blocks–or if you run later blocks price them lower to compensate on performance loses.

6. Listen to performance buyers.

7. Competitors are not a big issue–people buy multiple greens products/health products.

8. And please please please, support Podscribe’s effort in pulling this data.

Seriously, I love Podscribe and their team. This is a labor of care and attention to details.

If you don’t use Podscribe for your pixel measurement, ad verification, and planning… then keep doing what you’re doing. I don’t need more competition in this space. But, if you want to be a smarter buyer then check them out.

If you haven’t downloaded your copy, you can do so here: Q2 PPB Breakdown