Do Podcast Audiences Convert, or Only Recall?

Why is this so hard to answer?

Because of the structural gap documented elsewhere in this series: a podcast is delivered as a file to a device that, in the IAB's own words, cannot report back what was consumed. Nothing in the delivery path records that a person heard an advertisement, let alone acted on it.

Every attribution method in use is therefore an attempt to reconstruct a link that was never recorded. Promo codes ask the listener to identify themselves. Vanity URLs ask them to type something specific. Pixel matching infers a connection from network metadata. Each is a proxy, and each fails in a different direction.

This is unusual. Most digital advertising carries the click that connects exposure to outcome, which is why its measurement debates concern attribution models rather than whether a connection exists at all. Podcasting has no click.

So the honest answer to whether podcast audiences convert is that they demonstrably do sometimes — promo codes get redeemed, vanity URLs get typed — and that the rate cannot be established with any method currently available. What follows is an account of how each method fails, which is more useful than a number nobody can stand behind.

The gap is also why podcast advertising is often bought on faith and defended on anecdote. Where no instrument settles the question, the loudest evidence becomes a memorable success story, which is the weakest kind of evidence and the most persuasive in a meeting.

What do promo codes actually capture?

A minority of conversions, and a biased one. The most-cited figure comes from a vendor benchmark reportedly drawn from more than 66,000 campaigns, which found promo codes accounting for around 21% of total conversions attributed to podcast advertising.

If that figure is close to right, then a campaign measured by promo code alone is seeing roughly a fifth of its effect and drawing conclusions from it. An advertiser deciding that podcasts underperform on that basis is making a decision about a measurement artefact.

The bias matters as much as the undercount. A promo code captures listeners who were price-conscious enough to care about a discount, attentive enough to catch the code, and organised enough to remember it at checkout. That is a particular kind of buyer, and campaigns aimed at considered or high-value purchases will look worst under this instrument.

Timing compounds the problem. The same vendor research reports that roughly half of conversions occur more than a week after exposure, which places them outside the windows many campaigns use. A listener who hears a recommendation, thinks about it, and buys a fortnight later is invisible to most of this machinery.

Codes also decay as an instrument. A listener who hears one twice remembers it better, which means long-running campaigns measure better than short ones for reasons unrelated to effectiveness.

What do vanity URLs miss?

Everyone who behaves normally. The method depends on a listener hearing an address and typing it precisely, which is not how people generally respond to hearing about something interesting.

The common failure is well described in the trade literature: a listener hears a brand name with a slash-podcast suffix, and simply searches for the brand instead. The conversion happens, and it is recorded as organic search or direct traffic. The podcast gets no credit, and search gets credit it did not earn.

This produces a systematic misallocation rather than random noise. Audio advertising pushes demand toward search, where it is captured by whichever channel the buyer happens to arrive through, which makes brand search look effective and audio look weak — and encourages budget to move in exactly the wrong direction.

It also connects to a finding reported separately in this series. Branded search is a plausible downstream effect of coverage and a poor instrument for proving any single placement worked, because too many things move it at once. Here the same ambiguity reappears from the other side: the conversions are real, and no method reliably assigns them.

What does pixel matching add, and what does it assume?

It removes the demand that listeners do anything, and replaces it with an inference about network addresses.

The method compares the IP address recorded when an episode was requested with the IP address present at a later conversion event on the advertiser's site. Where they match within a window, the conversion is attributed to the exposure. No code, no typing, no memory required.

The assumption is that an IP address identifies a household or person reliably enough for the match to mean something. On a home connection that is often reasonable. On a shared office network, a university, a café or a mobile carrier's shared address space it is not, and the method will both invent matches between unrelated people and miss real ones when a listener hears an episode on mobile and buys on a laptop elsewhere.

So pixel matching is better at scale and worse at certainty. It produces a larger, smoother number with an error structure nobody outside the vendor can inspect, which is an improvement over promo codes and not the same thing as a measurement.

Who is producing these figures?

The companies selling the solution, almost without exception, and this is the qualification that should shape how the numbers are read.

The 21% promo code figure comes from an attribution vendor. The comparisons showing pixels outperforming codes come from firms selling pixel-based measurement. The articles explaining why promo codes undercount are published by companies whose product is the alternative. None of this makes the findings wrong — the mechanisms described are real and the reasoning is sound — but the direction of error is predictable, and a benchmark showing the incumbent method capturing a fifth of conversions is a benchmark with a commercial purpose.

Two figures come from a different kind of source and are worth separating. Survey work reports that only 49% of marketers describe themselves as extremely or very confident measuring podcast ad return, and that 33% of advertisers told the IAB that difficulty measuring performance is a direct barrier to increasing podcast spend.

Those are measures of professional confidence rather than of conversion, and they point the same way. The people buying podcast advertising do not believe they can measure it, and roughly a third say the uncertainty is capping what they spend.

What this means for anyone buying a podcast placement

Do not accept a single attribution number as the performance of a campaign. Ask which method produced it, what that method systematically misses, and what window was used. A promo code result presented without the caveat that codes capture a minority of conversions is a partial figure presented as a total.

Use more than one instrument, and expect them to disagree. A promo code, a vanity URL, a pixel where available and a post-purchase survey asking customers where they heard of you will produce four different pictures, and the spread between them is the honest confidence interval.

Extend the window. If roughly half of conversions arrive more than a week after exposure, a seven-day window is measuring the faster half and calling it the whole.

And be clear about what a guest appearance is, as distinct from an ad read. Media Morpheus books podcast appearances, which reach an audience and produce a citable record. Neither we nor anyone else can tell you what share of that audience will buy, because the medium does not record the connection and every method for reconstructing it fails in a way its vendor is not incentivised to emphasise.

Methodology and limitations

This report describes podcast attribution methods and their limitations as documented in trade literature, and reports figures as they were published rather than as verified measurements. Six limits apply, and the first is the most serious. First, the quantitative figures — the roughly 21% promo code share, the reported campaign count, the comparison between codes and pixels, and the timing of conversions — originate with companies selling podcast attribution products, and we could not retrieve the underlying benchmarks in primary form to assess sample construction or method. Second, the survey figures on marketer confidence and on measurement as a barrier to spend describe different populations (marketers and advertisers respectively) and are reported secondhand. Third, no independent academic or disinterested research on podcast conversion rates was located. Fourth, the comparison of attribution methods is our own ordinal classification of documented limitations, not a measurement of accuracy. Fifth, attribution practice changes quickly and this describes the position at the date of publication. Sixth, this report concerns advertising attribution; a guest appearance is a different product and is not measured by any of these instruments.

Conclusion

Podcast audiences plainly convert — codes get redeemed and URLs get typed — and nobody can say at what rate. The medium records no link between hearing something and doing something, so every method in use reconstructs that link from a proxy, and each proxy fails in its own direction: promo codes catch a price-conscious minority reportedly around a fifth of the total, vanity URLs lose everyone who searches instead of typing, and pixel matching trades certainty for reach by inferring identity from network addresses. Compounding this, almost every published figure comes from a firm selling the measurement. The most credible numbers in this area are the ones about confidence rather than conversion: roughly half of marketers do not believe they can measure this, and a third of advertisers say that is why they spend less.

Frequently Asked Questions

Sources

1. vendor research Podscribe benchmark analysis of podcast campaign attribution, as reported in trade coverage. Figures reported secondhand; primary benchmark not retrieved.
2. trade press Podcast attribution trade literature from measurement and agency firms describing the limitations of promo codes, vanity URLs and pixel matching.
3. trade body IAB, reported survey finding that 33% of advertisers cite difficulty measuring performance as a barrier to increasing podcast spend.