Podcast analytics come from two different kinds of counting. Your host counts downloads from its server logs, and IAB Tech Lab's podcast measurement guidelines treat a request as a valid download only once enough of the file for one minute of playback has been delivered. Apple Podcasts, Spotify and YouTube count what happens inside their own apps instead: plays, listeners, followers and time spent. The two families never match, and they are not supposed to.
This page defines the metrics you will meet on each dashboard, explains why they disagree, and ends with a monthly sheet for tracking your own show. Definitions were read from IAB Tech Lab, Apple Podcasts for Creators, Spotify for Creators and YouTube on 1 October 2026. Turning these numbers into a sponsor pitch is a separate job, covered by the media kit template.
Two families of numbers
Downloads exist because podcasting runs on open RSS feeds. An app requests the episode file from your host's server, and the host logs the request; whether anyone pressed play is usually invisible to it, because aggregator apps such as Apple Podcasts and Spotify typically do not send playback data back to publishers, as the IAB document notes. In-app metrics are the reverse. Apple, Spotify and YouTube see playback directly, but only inside their own apps.
IAB Tech Lab's announcement of version 2.3, finalized in September 2026, says the rules now apply explicitly to video episodes delivered through open RSS feeds, replace the word “listener” with “podcast consumer”, and stress that a download is not the same thing as a confirmed play. The same IAB post says a version 3.0 that moves beyond downloads is planned for public comment in the middle of 2027.
Cross-platform metric glossary
| Metric | Where you see it | How it is counted | Source |
|---|---|---|---|
| Download | Host dashboards | A filtered, de-duplicated request that delivered at least one minute of content, or the whole file for very short episodes | IAB v2.3 |
| Podcast consumer, formerly listener | Host dashboards | A unique pairing of IP address and user agent within a stated time frame | IAB v2.3 |
| Listeners | Apple Podcasts Connect | People who played or watched more than zero seconds of an episode | Apple Analytics |
| Engaged listeners | Apple Podcasts Connect | People who heard or watched at least 20 minutes or 40% of an episode | Apple Analytics |
| Average consumption | Apple Podcasts Connect | The average share of an episode played per device; above 100% means people replayed it | Apple Analytics |
| Plays on Spotify | Spotify for Creators | An episode watched or listened to for at least 30 seconds on Spotify | Spotify glossary |
| Spotify audience | Spotify for Creators | Distinct people who played any episode in the period, split into new and returning against the prior year | Spotify glossary |
| Completion rate | Spotify for Creators | Share of the audience that reached at least 95% of an episode in its first 7 days, with quartiles at 25%, 50% and 75% | Spotify glossary |
| Impressions and conversion rate | Spotify for Creators | How often your show or episodes were displayed on Spotify, and the share of those displays that led to plays | Spotify glossary |
| Followers | Apple and Spotify | Apple reports everyone following the show, recalculated daily; Spotify reports people who started following during the selected period | Apple follows, Spotify |
| Views and engaged views | YouTube Studio | Since August 24, 2026 a view counts from the first frame; an engaged view needs a click to watch or viewing past the first seconds | YouTube Blog |
| Watch time and retention | YouTube Studio podcast report | Time watched or listened, plus audience retention, traffic sources and demographics for the podcast playlist | YouTube Help |
One episode, several dashboards: worked scenarios
Applying the definitions to ordinary behavior shows why no single number is the real audience. Each row follows one hypothetical person.
| What the person does | Host dashboard | Platform dashboard | Why |
|---|---|---|---|
| A follower's Apple Podcasts app downloads the episode overnight, and it is never played | Counts a download | No Apple listener | Delivery happened; playback never did |
| Someone taps play in Apple Podcasts and stops almost at once | May not count, if too little of the file was fetched | Counts an Apple listener but not an engaged one | Apple needs any playback; the host needs a minute of content delivered |
| A Spotify user listens briefly, then leaves | Often nothing, because Spotify normally streams its own copy | Not a Spotify play | Spotify plays need a minimum listen; files are served from Spotify unless the host uses RSS passthrough |
| A thumbnail autoplays on someone's YouTube home page for a moment before they scroll on | Nothing, since YouTube streams its own upload | A view but not an engaged view | YouTube's view now starts at the first frame |
| A listener on a small app streams the episode in two sessions a day apart | One or two downloads | Not visible to Apple, Spotify or YouTube | The host's measurement window decides |
The Spotify row needs the most explanation. Spotify's podcast delivery specification says Spotify downloads every episode from your feed and normally handles playout itself, with an RSS passthrough option for some providers in which the Spotify app streams directly from the source. Unless your host has that arrangement, Spotify listening may barely show up in your download counts, so use Spotify for Creators for Spotify numbers and ask your host which setup applies.
Why the dashboards disagree
- Different thresholds. A brief listen can count as an Apple listener, fall short of a Spotify play and still miss a host's download threshold, because each platform draws its line in a different place; the glossary above links every definition.
- Downloads are not followers. Apple's follow guide lists the reasons: people play without following, followers download on several devices, some switch automatic downloads off, and Apple pauses downloads for followers who stop listening.
- Counting windows. IAB's version 2.3 describes three ways of de-duplicating repeat requests (a fixed calendar day, a rolling window from the first request and a rolling window that resets with each request) and says differences within ±5% between certified measurers are generally acceptable.
- Host moves. IAB notes that a changed enclosure URL can make apps download an episode again, and Apple's hosting change guide warns host dashboards may look atypical after migrating while Apple's own analytics stay consistent.
- Thresholds and delays. Apple's data requirements hide figures until at least five unique users have played the show in the selected period and say some data can arrive up to 72 hours late, while Spotify's analytics pages say most figures refresh every 15 minutes and some take up to 48 hours.
Spotify adds one more wrinkle for shows that switch to its hosting: its data changes article explains that other platforms report downloads while Spotify reports plays, and that its analytics begin on the day you switched.
What IAB certification tells you, and what it does not
A host on IAB Tech Lab's compliant companies list has had its method for counting downloads and podcast consumers audited against the guidelines, and some are also certified for ad delivery. That gives sponsors a shared basis for comparing shows that use different hosts. It does not tell anyone whether people finished the episode, who they are, or whether an ad was heard, since a delivered file is not a confirmed play. The entries on the list when it was read were certified against version 2.2. Whether your host appears there is part of the podcast hosting comparison.
Reading drop-off and completion
Consumption curves are among the most useful numbers for improving the show itself. Apple's Analytics overview frames them as questions, such as whether listeners skip your intro or your second mid-roll, and its Performance tab compares an episode with your median, your average or your best episode over its first weeks. Spotify's overview article shows whether an episode sits inside your normal range once you have 10 episodes published. YouTube's podcast report adds retention for the playlist's videos.
- Compare each episode with your own median, never with another show's public numbers.
- Look for steep drops at the intro, at ad breaks and at segment changes, and note the timestamp.
- Check what the script said at that point and change one thing in the next episode.
- Wait a week before judging a new episode, since early numbers move quickly.
The podcast script template marks segments and breaks in a way that makes those timestamps easy to match.
Monthly podcast KPI sheet
Copy this into a spreadsheet and fill it on the same day each month. The definition column stops you comparing unlike numbers; the last column forces one decision.
| KPI | Where to pull it | Definition to keep in mind | This month | Last month | What you will change |
|---|---|---|---|---|---|
| Downloads per new episode, first month | Host dashboard | Server-side, IAB-style counting | |||
| Net new followers | Apple Trends and Spotify Engagement | Apple totals versus Spotify new follows | |||
| Engaged listeners | Apple Podcasts Connect | Apple's engaged rule from the glossary | |||
| First-week completion | Spotify for Creators | Spotify-only, new episodes | |||
| Impression to play conversion | Spotify Discovery tab | Spotify surfaces only | |||
| Average view duration | YouTube podcast analytics | Engaged views anchor this figure | |||
| Biggest drop-off point | Apple, Spotify or YouTube curves | Timestamp plus script block | |||
| Top discovery source | YouTube traffic sources, Spotify discovery | Platform-specific labels | |||
| Episodes published | Your release calendar | Context for every other row |
Export the raw data when you fill the sheet, because dashboards change and old periods are easy to lose. Spotify offers CSV export from the web version, and the Spotify for Creators setup guide shows where each analytics tab lives; YouTube's side is covered in the guide to podcasts on YouTube.
Limitations
Every definition here can change, and two changed in the weeks before this page was written, when YouTube redefined views and IAB finalized version 2.3. No benchmark figures are offered, because what counts as a healthy number depends on topic, format and audience, and published averages often do not say how they were measured. Listening in apps that report nothing back to you, private feeds and shared devices all blur the picture. Analytics also arrive aggregated for privacy; Apple, for example, states that its data contains no personally identifiable listener information. Treat these numbers as directions for your next episode, not verdicts on the show.