How to Spot Inflated Podcast Downloads in 2026

How to Spot Inflated Podcast Downloads in 2026

Updated October 2026 · Based on CastFox data covering 28 Apple Podcasts chart snapshots in the United States (2026-02-24 to 2026-10-13) across business, health and fitness, and technology, 758 distinct shows that passed through those three top 100 charts, and country breadth measured over the 20 to 22 snapshots per genre where we recorded all 78 countries.

You cannot audit a podcast's downloads from outside, and in 2026 you do not have to in order to catch the obvious cases. A download figure is produced by the show's own hosting platform and handed to you in a media kit, so the way to test it is to check it against signals the show does not control. In our chart data, the 123 shows that held a place in all 28 US snapshots across three genres charted in a median of 37 countries at once and had published a median of 372 episodes. The 164 shows that appeared in exactly one snapshot charted in a median of 2 countries and had published 97.

That gap is the whole method. A real audience leaves a trail in places the seller has no access to: other countries' charts, a catalog that took years to build, a publishing cadence, repeat sponsors. An inflated number leaves a trail in exactly one place, the report the seller sends you. When a show claims reach that would put it in the top tier of a genre and none of the external signals agree, you are not looking at proof of fraud, you are looking at a reason to ask for a certified report before you wire anything.

This post sets out what podcast download fraud looks like in practice, what our chart data says a durable audience actually looks like in three genres, the six checks we would run before any sponsorship, and the questions to put to a show whose numbers do not line up. Every CastFox figure here is chart and catalog data. We do not measure downloads, and nothing in this post should be read as an accusation against any named show.

Why Download Numbers Cannot Be Checked at Source

A podcast download is a server event. Someone's app requests the audio file, the hosting platform logs the request, and at the end of the month the platform tells the show how many requests passed its threshold. The advertiser is three steps removed from that log and has no way to read it. There is no equivalent of a media auditor who can open the books, and the figure you are quoted has usually been copied into a slide by the person selling you the slot.

The industry's answer is measurement certification: hosting platforms submit their counting rules to an independent technical audit, and a certified platform applies a shared definition of what counts as a download, filters known bot traffic, and excludes requests that did not fetch enough of the file. That certification is worth insisting on, and it is the first question in our checklist below. It is not a complete answer, because it certifies the platform's counting method rather than the specific numbers in a media kit, and because a show can produce genuine requests from an audience that was paid to generate them.

So the practical approach for a buyer is triangulation. Ask for the certified report, then test the story it tells against evidence the show cannot write: whether the audience shows up on charts in other countries, whether the catalog is old enough to have accumulated that audience, whether the publishing cadence matches a show with an engaged following, and whether any other advertiser has come back for a second flight. None of these is a measurement of downloads. Together they tell you quickly whether a claimed audience is plausible.

Podcast Download Fraud: What It Actually Is

Podcast download fraud covers a spread of behavior, and most of what advertisers meet is at the mild end. In rough order of how often buyers encounter it:

• Counting generously. A show reports every request rather than the ones that fetched enough audio to count as a listen, or reports a lifetime figure as a per episode average, or quotes its best month as typical. No bots involved, and the number is still two or three times too high.
• Mixing metrics. Video views, social impressions, newsletter subscribers, and audio downloads get added together into one "reach" number. This is the most common form of inflation in a media kit and the easiest to unpick, because the components are usually visible in public.
• Prefetch and auto download. Apps that download new episodes automatically for subscribers generate requests whether anyone listens or not. A certified platform applies rules to this; a self reported number often does not distinguish it.
• Bought traffic. Paid services generate requests against the episode file from data centers or device farms. Certified platforms filter known sources, which is why the practice shows up more often on shows that cannot produce a certified report.
• Incentivized listening. Audiences paid or rewarded to subscribe and let episodes play. The requests are real and from real devices, so filtering does not catch them, and the only external tell is that the audience never behaves like an audience: no chart movement outside one market, no reviews in proportion to the claimed size, no engagement anywhere else.

The reason to separate these is that they call for different responses. Generous counting and mixed metrics are negotiating problems, and you fix them by asking for the specific definition behind the number. Bought and incentivized traffic is a fraud problem, and you avoid it by refusing to buy on self reported figures alone.

What a Durable Audience Looks Like in Chart Data

CastFox records the Apple Podcasts top 100 in 78 countries every day. To build a benchmark we took every show that appeared in the US top 100 of three genres across our 28 snapshots between 2026-02-24 and 2026-10-13: business, health and fitness, and technology. That is 758 distinct shows. Then we split them into the shows that were there in every single snapshot and the shows that appeared in exactly one, and compared what each group looks like from outside.

Two things in that table matter to a buyer. The first is that a genre top 100 is much less stable than it looks: between 17% and 24% of all the shows that passed through these charts were there for a single snapshot, while only 40% to 42% of today's top 100 is held by shows that were present throughout. A screenshot of a chart position proves a day, not an audience. The second is that the two groups look different in ways that have nothing to do with the chart itself.

The country column is the one we would lean on hardest. Charting in many countries at once is difficult to manufacture, because it requires simultaneous listening in markets where a buyer of traffic has no particular reason to buy. 70% of the always present shows reached 10 or more countries at once against 21% of the one snapshot shows, and 42% of the one snapshot group never charted outside a single country against 1% of the durable group. Technology is the clearest case, where the median always present show peaks at 56 countries and the median one snapshot show at 7. In business the medians are 7 and 1, and in health and fitness 28.5 and 2.

The catalog column is a plausibility check rather than a fraud signal. Audiences accumulate. A show with a median 372 episode catalog has had years of weekly publishing to build the number it claims, and 41% of the one snapshot shows had published fewer than 50 episodes at all. A new show can be genuinely popular, but a new show claiming an audience that would place it in the top tier of its genre is a claim worth a certified report.

One caveat we want on the record, because it cuts against the neat version of this story. Country breadth on its own can spike: in the business sample, one show reached 68 countries in a single snapshot while appearing in only 2 of our 28 US snapshots, which is what a coordinated launch or a network promotion looks like rather than anything improper. The signals work as a set, not individually.

How to Verify Podcast Download Numbers

Verification here means two separate jobs: establishing that the reported number was produced by a defensible counting method, and establishing that the audience it describes behaves like a real one. Do them in that order, because the first is a single email and the second takes about five minutes per show.

Ask for the platform report, not the media kit figure

Request a report exported from the show's hosting platform, covering the last 90 days, broken out by episode, and state that you need the platform to be one with a recognized measurement certification. Ask for it as a document rather than a number in an email. A show selling real inventory produces this in a day, because it is three clicks in their dashboard. A show that cannot or will not is telling you something before you have looked at any data.

Read unique listeners before downloads

In the same report, look for 30 day unique devices or unique listeners and put it next to the download count. The ratio between them is the shape of the audience. A show with a large download number and a small unique count is either serving a small audience repeatedly or counting requests that are not people, and in either case the reach you are buying is smaller than the headline. Ask for the ratio explicitly if the report does not show it.

Cross check the audience against charts in other countries

This is the check the seller cannot influence. Look up the show's chart history and count the countries it has charted in and the number of days it held position. Our benchmark above gives you the comparison: a median of 37 countries and continuous presence for durable shows against 2 countries and a single appearance for the churn. A show claiming a top tier audience that has never charted outside its home market needs an explanation, and there are legitimate ones, such as a show in a language spoken in one country, so ask rather than assume.

Check the catalog, the cadence, and the last episode date

Open the feed and look at three things: how many episodes exist, how regularly they ship, and when the most recent one landed. A gap of several weeks in a show claiming a weekly audience is the most common quiet problem in a media kit, because the figure being quoted predates the gap. You can see all of this on the show's page in CastFox analysis alongside its chart record.

Ask who else has bought, and who bought twice

Ask for two or three brands that have run on the show more than once, and ask for the promo codes they used. Repeat buyers are the strongest signal available to you, because another advertiser has already measured the show and chosen to come back. A show with a long list of one time sponsors and no repeats has been tested and quietly failed, whatever the download figure says.

Signs a Podcast Is Buying Downloads

No single item below proves anything. Two or three together, in a show that will not produce a certified report, is where we would walk away.

Reach that does not match the chart record

A claimed audience in the hundreds of thousands per episode, from a show that has never held a genre top 100 position for more than a snapshot or two. In our three genres, every show that charted continuously for all 28 snapshots also charted in more than one country; 1% of that group never did. A big number with no chart footprint anywhere is the clearest mismatch in the set.

Traffic concentrated in a single market

Ask for the geographic breakdown in the platform report. Bought traffic tends to arrive from a narrow set of locations, and so does incentivized listening. A show claiming a large international audience whose report shows one country, or whose chart history shows one country, is worth a direct question.

A download curve with no tail

Real episodes keep being downloaded for weeks. In the report, look at downloads in the first 7 days against the figure at 30 and 90 days. An audience that stops the moment a campaign ends, with almost no accumulation afterward, does not look like a subscriber base.

A claimed audience much larger than the catalog supports

The median always present show in our sample had published 372 episodes against 97 for shows that appeared once, and 41% of the one snapshot group had fewer than 50 episodes. A show with 20 episodes and a top tier download claim is not impossible, it is just unlikely enough to need the report.

Engagement that does not scale with the claim

Reviews, ratings in multiple countries, social following, and comment volume all scale roughly with audience size within a genre. A show claiming a six figure per episode audience with a handful of ratings and no measurable engagement anywhere else is describing listeners who do nothing but download, which is the signature of traffic that was purchased rather than earned.

A media kit with one number and no definitions

No methodology line, no platform named, no date range, no distinction between downloads and views. This is the cheapest tell of all, and it is usually carelessness rather than fraud. Either way, you cannot buy against it, so send the request for a certified report and see what comes back.

Podcast Audience Verification for Advertisers

Here is the sequence we would run on any show before a first flight. Three of the six checks need nothing from the seller, which is why we would do those first: if a show fails them, the email asking for a report never has to be sent.

Weight the checks by what you are risking. For a test flight of a few thousand dollars, the three external checks plus a plain request for the platform report is proportionate. For an annual commitment or a category exclusive, insist on the certified report with unique listeners and the geographic breakdown, and ask to speak to a repeat advertiser. The cost of the extra diligence is one email thread, and the cost of skipping it is the whole flight.

Shows That Clear Every External Check

To make the benchmark concrete, these are shows from the three genres we sampled that pass all three seller independent checks: present in every one of our 28 US snapshots, charting in a wide set of countries at once, and publishing a catalog deep enough to support a large audience. This is not a ranking and it is not a buy list. It is what the external signals look like when they all agree, so you have something to compare a media kit against.

Nothing here says what these shows cost or how many downloads they get, because we do not measure either. It says that if one of them sent you a large reach claim, the outside evidence would already be consistent with it. For the full data driven rankings in each category see best business podcasts, best health and fitness podcasts, and best technology podcasts.

How to Vet a Show Before Sponsoring or Pitching It

The verification checks above fit inside the workflow CastFox is built around. These four steps take about twenty minutes per show and work the same whether you are buying an ad or pitching a guest.

Discover: start from the chart record, not the media kit

Filter by genre and country and build the candidate list from shows that hold position across snapshots rather than shows that are high today. In our three genres that is 40% to 42% of the current top 100, and it is the part of the list whose audience you can be confident still exists next quarter. CastFox discovery shows the chart history next to each name, so a show that appeared last week is obvious before you contact it.

Analyze: check the show against the claim

Open the show and read the three external signals together: snapshots present, countries charted, catalog and cadence. Then look at which brands have run ads on it and which have come back. CastFox analysis puts the chart record, the catalog, and the detected sponsors on one page, which is where the mismatch with an inflated media kit shows up.

Organize: keep the evidence with the shortlist

Record what you verified and when, next to the show. Download claims change every quarter and so do publishing schedules, so a shortlist without dates on the evidence is a shortlist you have to rebuild. CastFox lists hold the candidates, the chart data, and your notes together.

Pitch: make the report part of the first ask

Put the request in the opening email: a 90 day platform report by episode, from a certified host, with unique listeners and the country breakdown. Asking early is not adversarial, it is standard, and it sorts the shows that sell real inventory from the ones that will spend three weeks not answering. CastFox pitching keeps the ask and the follow up in one place.

Methodology: Where These Numbers Come From

CastFox records the Apple Podcasts top 100 in 78 countries every day. For this post we used 28 US snapshots between 2026-02-24 and 2026-10-13 in three genres: business (249 distinct shows through the top 100), health and fitness (286), and technology (223). A show is counted as always present when it appears in all 28 US snapshots for its genre, and as a one snapshot show when it appears in exactly one. Episode counts and publishing frequency come from the show's own feed metadata.

Country breadth is the largest number of countries in which a show was simultaneously inside that genre's top 100 on a single snapshot date. Our country coverage per date varies, so this figure is calculated only over the dates where we recorded all 78 countries: 22 of 28 dates in business, 20 in health and fitness, and 22 in technology. Measuring breadth on a single shared date would have scored every show not charting that day as zero countries, which is why we take each show's own peak instead.

Three limits to state plainly. CastFox does not measure downloads, impressions, or spend, and no number in this post is a download figure: the comparisons here are chart, country, and catalog data used as plausibility checks on claims made elsewhere. The cohort comparison describes groups, not individual shows, and a show appearing in one snapshot is not evidence of anything improper, as new launches, seasonal shows, and network promotions all produce the same pattern. And the three genres sampled are large English language categories; a smaller or non English genre will have different medians, so treat the direction of the gap as the finding rather than the exact figures.

Frequently Asked Questions

How to check if a podcast audience is real

Run the three checks the show cannot influence before you ask it for anything. Count how many days or snapshots it has held a genre chart position, count how many countries it charts in at once, and look at the episode catalog and the date of the most recent episode. In our sample of 758 shows across three genres, the ones that charted continuously reached a median of 37 countries with 372 episodes published, against 2 countries and 97 episodes for shows that appeared once. Then ask for a 90 day report from a certified hosting platform showing unique listeners next to downloads.

What is podcast download fraud?

Any practice that makes a show's reported audience larger than its real one. In practice it runs from generous counting and mixing video views into an audio number, through requests generated by data centers or device farms, to audiences paid to subscribe and play episodes. The last of these is hardest to detect because the requests come from real devices, which is why external signals such as chart breadth matter more than any single report.

Can an advertiser verify podcast downloads independently?

Not directly. Downloads are logged by the show's hosting platform and there is no third party an advertiser can query for a second opinion. What you can do is insist the figure comes from a platform with an independent measurement certification, read unique listeners alongside downloads, and test the claimed audience against evidence the show does not control. CastFox supplies the third part: chart history across 78 countries, catalog and cadence, and the brands detected running ads on the show.

Is a short chart appearance a red flag on its own?

No. Between 17% and 24% of all the shows that passed through the three genre top 100s we sampled were there for a single snapshot, and most of them are ordinary: new launches, a network promotion, a guest who brought an audience for a week. It becomes a red flag only when it sits next to a large reach claim, no presence in any other country, and an unwillingness to produce a platform report.

What should I ask a podcast for before sponsoring it?

Four things, in one email. A 90 day report exported from the hosting platform and broken out by episode, the platform's name and its measurement certification, 30 day unique listeners next to the download figure, and the geographic breakdown. Add a fifth if the commitment is large: two brands that have run on the show more than once. A show selling real inventory sends all of this within a day.

Do inflated downloads matter if the campaign still performs?

Yes, because you are paying on the inflated number. A CPM negotiated against a reach figure that is twice the real audience is a rate twice what you agreed to, and the attribution you collect afterwards measures the small real audience rather than the one you bought. If a show performs well against a verified small audience, that is the better buy, and it is the argument we made in why a targeted audience beats wide reach.

Related Reading

• How to Vet a Podcast Before Sponsoring It: 2026 Checklist
• Podcast Sponsorship Rates by Audience Size 2026
• Stop Chasing Downloads: Why Audience Fit Beats Audience Size
• How Podcast Audience Demographics Are Calculated
• How Engagement and Reviews Predict a Podcast's Real Audience
• Most Consistent Podcasts of 2026
• Podcast Media Buying Guide 2026
• Podcast Advertising CPM by Genre 2026
• How to Measure Podcast Advertising ROI in 2026
• How to Benchmark a Show's Audience Before You Buy

Explore CastFox

• Best Business Podcasts and Best Technology Podcasts
• Best Health and Fitness Podcasts
• Analyze a show's chart record and sponsors
• CastFox for advertisers
• Compare shows side by side