IAB Australia today published the Audio Buyer's Guide Glossary 2026, a 10-page reference compiled by its Audio Advertising Council that defines 55 terms across six categories, from RSS feeds to unreserved fixed rate deals. Read closely, the glossary sets at least four different time thresholds for counting audio activity - one second, two seconds, 60 seconds and a 24-hour window - and three different ways of defining a month.

In Short

The Australian ad industry body wrote a dictionary of the words people use when buying ads in podcasts and streaming audio. It matters because the words decide what gets counted, and the dictionary counts a "play" after one second but an "active stream" only after a full minute. Anyone comparing audio numbers from different sellers now has a single list to check which of those definitions a report is using.

What IAB Australia published

The announcement page on the IAB Australia website is dated October 15, 2026, and filed under Research & Resources. According to IAB Australia, the glossary is "a reference for the key terms, formats and metrics used when planning, buying and measuring digital audio." The page says it was compiled by the Audio Advertising Council "as a companion to the Audio Buyer's Guide," and that it "covers everything from podcast fundamentals and dynamic ad insertion through to listening metrics and pricing models."

Two versions of the document exist. The web page is short: an introductory paragraph, the six category names, and a download button. The downloadable PDF is the substance - a 10-page A4 document produced in Adobe InDesign, with a cover, a contents page, eight pages of definitions and a closing page that carries only the running footer. The file's metadata shows it was created on October 14, 2026, at 22:15 UTC, the evening before the web page went live.

The two versions agree on the six category names and their order: Audio Fundamentals, Ad Formats & Delivery, Audience, Data & Targeting, Audience & Listening Metrics, Ad Delivery & Measurement, and Buying & Pricing. Both state that terms are ordered alphabetically inside each category. No term or definition appears on the web page, so no definitions conflict between the two.

One discrepancy sits on the PDF's own contents page. While every page footer reads "audio buyer's guide | glossary | 2026", the contents page states that the glossary was "Compiled by the IAB Australia Audio Advertising Council for the Audio Buyer's Guide 2025." The web page drops the year and refers only to "the Audio Buyer's Guide." Whether the 2026 glossary is a refreshed edition of a list first written for a 2025 guide, or a holdover line left in the layout, is not stated in either version. The naming of the council also varies across IAB Australia material. PPC Land's coverage of the tenth Audio Advertising State of the Nation report in March 2026 and of two alcohol advertising compliance guides in August 2025 both refer to the body as the "Audio Council", which was set up in 2015.

How the 55 entries break down

The categories are uneven. Audience & Listening Metrics is by far the largest, with 19 entries spread across two pages. Audio Fundamentals holds 10, Audience, Data & Targeting holds eight, Ad Formats & Delivery and Ad Delivery & Measurement hold seven each, and Buying & Pricing holds four.

That distribution says something about where the council sees confusion. Nearly half the document - 26 of 55 entries - deals with counting: how many people listened, for how long, and how many ads were delivered or heard. Only four entries address price.

Several entries are not separate concepts at all. Time Spent Listening (TSL) is a pure cross-reference: its definition reads, in full, "See Average Time Spent Listening (ATSL)." Plays and Session Starts (SS) carry word-for-word identical definitions - "A user presses play on content and listens for more than one second" - each pointing to the other. Completion Rate and Listen Through Rate (LTR) are likewise identical: "The percentage of an ad that a listener engages with." Strip out duplicates and cross-references and the glossary describes 52 distinct ideas.

Four thresholds, one industry

The most practical content in the document is not any single definition but the thresholds buried inside them.

At the bottom sits one second. Under the glossary, a play or a session start counts once a user "listens for more than one second." Next comes two seconds: an Audible Ad Impression "excludes ads audible for less than two continuous seconds," and according to the glossary, "Any part of the ad can meet the two-second threshold."

Then the bar jumps to a minute. An Active Stream is "Listening to content for 60 seconds or longer." Bounce Rate (Audio) is the "Percentage of streaming sessions that last less than 60 seconds." Average Time Spent Listening only counts "sessions over one minute." For streaming, Reach is "the number of unique IP addresses that listened to a station for at least one minute in a day." And a podcast Download requires that "at least one minute (60 seconds) of podcast content must be downloaded by a unique IP address and user agent within a 24-hour period."

The result is a 60-fold gap between the lowest and the most common qualifying threshold. A streaming service reporting plays and one reporting active streams, both using IAB Australia's own vocabulary, can describe the same session in numbers that cannot be compared. The glossary does not rank these metrics or say which should be used for buying. It sets them side by side.

Ad-level measurement adds a different scale. A Client Confirmed Ad Play is confirmed "by a tracking beacon sent from the client (the listening app or device) when the file plays," and according to the glossary it is "reported at first quartile (25%), midpoint (50%), third quartile (75%) and complete (100%)." A Delivered Ad Impression (Podcast), by contrast, is determined "by server logs, when either all bytes of the ad file were sent or the bytes of the podcast file containing the ad were downloaded." The first describes what the device played; the second describes what the server sent. Neither confirms that anyone heard the ad, which is the gap the two-second audible impression tries to fill.

Three ways to define a month

Time windows are no more consistent. Average Monthly Listener is the "Average number of unique users who have listened at least once within a 28-day period." An Active Signed Up User is "An active user (last 30 days)" who has provided date of birth and gender. Monthly Active User (MAU) is "The average number of users who open the app at least once a month."

So a month is 28 days in one entry, 30 days in another, and an undefined calendar month in a third. The week is simpler: Average Weekly Listener uses a seven-day period. Daily Active Listeners is "The number of listeners who listen to a stream each day," without a minimum duration.

There is also a quiet difference between opening an app and listening. MAU and the broader Active Users (AU) family - daily, weekly and monthly - count users "who open the app." The listener metrics count people who actually listened. A publisher reporting MAU is therefore reporting a figure that, under these definitions, does not require a single second of audio.

People, devices and IP addresses

The glossary uses three different units for an audience and does not always say which one a metric counts.

A Listener is "a single person or user." Unique Listeners, however, are the "Number of unique active devices listening, determined by a combination of IP address and user agent (device)." Reach is "The total number of people who interact with content" - and then, for streaming, is redefined as a count of "unique IP addresses." People, devices and IP addresses are not interchangeable. A household sharing one router can appear as one IP address, while a commuter whose phone moves between mobile networks can appear as several.

This is not a flaw invented in Australia. When IAB Tech Lab finalised its podcast rules, PPC Land reported that the v2.3 Podcast Technical Measurement Guidelines acknowledge mobile "IP-hopping" can distort counts and truncate IPv6 addresses to their first 64 bits. The glossary does not mention either issue.

Where the glossary lags IAB Tech Lab

The Download entry ties the Australian definition to "the IAB Tech Lab's Podcast Measurement Technical Guidelines," without naming a version. That matters because the guidelines changed weeks ago.

IAB Tech Lab put version 2.3 out as a 32-page draft on July 21, 2026, and finalised it on September 29. Three of the changes reported by PPC Land bear directly on the Australian text. First, v2.3 replaced "listener" with "podcast consumer" to cover people who watch as well as hear. The glossary still uses "listener" throughout. Second, v2.3 brought video episodes delivered through open RSS feeds, including HLS streams, under the same counting rules. The glossary contains no entry for video. Third, v2.3 names three measurement windows - calendar day, first touch and last touch - and requires only that the chosen one be applied consistently. The glossary describes a single "24-hour period."

The one-minute download floor and the IP-plus-user-agent rule match the Tech Lab text. So does the requirement that spiders and bots be removed, which the glossary states in a single clause. The wider category of invalid traffic gets one mention, in the Average Connected Streams entry, where "Data must be filtered to exclude invalid traffic." There is no separate entry for it.

The Tech Lab work on podcast counting has a long history. PPC Land covered the launch of the Podcast Measurement Compliance Program in December 2018, when NPR and Blubrry were the first to earn seals. The Australian glossary does not mention certification.

Podcasts are longer than 10 minutes

The glossary's definition of a Podcast is narrower than common usage. According to the glossary, a podcast is "On-demand audio content longer than 10 minutes that can be downloaded and listened to via the internet." Daily news briefings and short-form shows under 10 minutes fall outside the definition as written. The document does not say how such content should be classified or whether ads within it count as podcast advertising.

The same entry describes podcasts as "usually offered by subscription, with new episodes delivered automatically to subscribers." Subscription gets its own entry: "When a listener opts in to receive new podcast episodes automatically in their preferred podcast app." Neither describes paid subscriptions, which several podcast platforms now sell.

The definition of Aggregator names Apple Podcasts, Spotify and YouTube Music. Listing a YouTube property while the rest of the document stays silent on video is the clearest sign that the glossary describes audio as it was bought a few years ago rather than as it is distributed now. Apple began testing HLS video podcasts with dynamic video ad insertion in February 2026, with Acast, ART19, Triton's Omny Studio and SiriusXM as launch hosting providers.

Streaming, downloads and the space between

The glossary handles the technical distinction between streaming and downloading with some care. Digital Audio is defined as "Any audio streamed or downloaded on a device connected to the internet," with the observation that "The technical difference between streaming and downloading is almost invisible to the listener, but it affects how advertising campaigns are planned, bought and measured."

Progressive Download addresses the confusion directly. According to the glossary, "Many podcast apps have a 'play now' function that feels like streaming to the user. Technically, the content (including the ad) is still delivered by progressive download: it downloads while the user listens." That explains why podcast impressions are counted from server logs rather than from playback, and why a downloaded ad is not a heard ad.

Dynamic Ad Insertion (DAI) carries two definitions in one entry. For streaming, it means "intercepting traditional radio broadcast streams and inserting audio ads into ad breaks as they play." For podcasts, it means "inserting ads into podcasts at the time of download." The glossary adds that "DAI is a form of ad serving." Ad Markers are the "Points chosen in a podcast episode where ads are dynamically inserted."

The opposite approach is Embedded Ads, "Also known as 'baked-in' ads." According to the glossary, these are "Ads built into the original podcast file at production. Every listener hears the same ad regardless of when they download the episode, and the publisher cannot update or replace it without re-editing the audio file." The baked-in model and DAI cross-reference each other, which is the only pair in the format section to do so.

Formats it includes, and the one it omits

Beyond pre-roll, mid-roll and post-roll placement, the format section lists three newer types. Companion Banner describes a display banner shown in the app alongside the audio ad, with the caveat that "It only appears if the user has the app open while the ad plays." Dynamic Audio Creative (DCO Audio) describes "Data-driven, personalised audio ads that change parts of the script based on real-time data such as time, day, weather and location." Voice Activated Ads let listeners respond by voice, "such as having information emailed or texted to them, or opening a website," and the glossary notes that "The listener needs a microphone connected to the playback device."

The format most associated with podcast advertising is missing. There is no entry for host-read ads, in which the presenter reads the advertiser's message in their own voice. Embedded ads are described purely in terms of how they are delivered, not who reads them. For a buyer's glossary, the omission is notable: host-read spots are a large part of what podcast sellers price and sell, and the distinction between a host-read spot that is baked in and one that is dynamically inserted is exactly the kind of question the document is meant to settle. Nor does the glossary define programmatic audio, sponsorship or branded podcasts.

Data and identity without privacy

The Audience, Data & Targeting section defines first-, second- and third-party data, cookies and device IDs. Addressable Advertising is inventory that carries "an identifier, such as a cookie, device ID or hashed email (HEM), that can be used to target advertising."

The Device ID (DID) entry describes a mobile advertising ID as "A unique identifier assigned to a mobile device and used for advertising in apps, where cookies are not available," adding that "Device IDs can be matched with existing user data (e.g. age or gender) without directly identifying individuals." Separate entries identify the IDFA as Apple's device ID and the AAID as Google's.

None of these entries mentions consent, opt-outs, Apple's App Tracking Transparency prompt or the availability of IDFAs after users decline tracking. The Cookie entry says cookies "were originally designed to improve the user experience" but does not discuss browser restrictions. The section reads as a description of the identifiers themselves, not of the conditions under which they can be used. The claim that device IDs can be matched with age or gender "without directly identifying individuals" is presented without qualification.

Third-party data receives a one-line definition: "Aggregated data from various external sources." It is one of the shortest entries in the document.

Four ways to pay

The Buying & Pricing section is the thinnest. It defines Cost Per Click (CPC), Cost Per Lead (CPL) and Cost Per Mille / Thousand (CPM) - the last as "total advertising spend divided by total impressions delivered, multiplied by 1,000." Click and lead pricing sit oddly in an audio context, where most ads carry no clickable element unless a companion banner is on screen.

The fourth entry, Unreserved Fixed Rate (UFR), is "A programmatic deal type where the buyer and seller agree a fixed price, with no guaranteed volume of inventory." It is the only programmatic deal type defined. Open auctions, a private marketplace and programmatic guaranteed deals are absent, even though the glossary includes the word "programmatic" in that single definition. There are no entries for cost per completed listen, cost per download or sponsorship rates.

Measurement of outcomes is also missing. The glossary has no entries for attribution, brand lift, incrementality or market mix modelling.

Why the vocabulary matters in Australia

The market the glossary describes is not small. PPC Land reported that Australian digital audio ad spend reached $339 million in calendar 2025, up 8.2%, with podcast advertising up 13.5% to $134 million and streaming audio up 5% to $205 million. For the financial year to June 30, 2026, audio advertising grew 6.6% to AU$353 million, with podcasting accounting for about 40% of audio spend and posting a quarterly record of AU$40 million in the June quarter.

The same State of the Nation report, published by IAB Australia in March, is the strongest argument for a shared vocabulary. Of 128 buy-side respondents surveyed in late 2025, 27% cited a lack of standardised audience data for digital audio planning, and 29% named fragmented audio formats as the top barrier to holistic measurement. A glossary is the cheapest possible response to those complaints. Whether it resolves them is another matter: it records the inconsistencies rather than choosing between them.

Measurement doubts are not confined to Australia. A Bauer Media Audio study covered by PPC Land found that only 13% of European advertisers confidently use audio attribution tools. An AdsWizz report found audio takes 31% of consumers' media time but receives 9% of ad budgets. Industry bodies have argued for years that shared definitions are a precondition for closing such gaps.

The more immediate pressure is video. A VAB report in September found that 88% of weekly US podcast consumers use video platforms, and Triton Digital's chief revenue officer had already named video discovery as one of four trends for programmatic audio in 2026. IAB Tech Lab moved its counting rules to cover video weeks before today's release. A buyer's glossary that treats podcasts as audio-only files, ordered for an RSS-era workflow, describes part of the market Australian buyers are now purchasing - but not all of it.

What the document does well is narrower and still useful. It places the arithmetic of audio measurement in one place: the one-second play, the two-second audible impression, the 60-second stream, the 28-day listener. Anyone reading a publisher's audience claim can now check which of those floors sits underneath it.

Timeline

Summary

Who: IAB Australia, through its Audio Advertising Council, published the glossary. It is aimed at media agencies, brands and publishers that plan, buy, sell and measure digital audio advertising in Australia.

What: The Audio Buyer's Guide Glossary 2026, a 10-page PDF defining 55 terms in six categories, accompanied by a short web page. The document sets thresholds ranging from one second for a play to 60 seconds for an active stream and a download, uses 28-day, 30-day and calendar-month windows, duplicates two pairs of definitions, and omits host-read ads, video podcasts, most programmatic deal types and attribution. Its PDF says it was compiled for the Audio Buyer's Guide 2025.

When: The web page is dated October 15, 2026. The PDF file was created on October 14, 2026.

Where: Australia, published on the IAB Australia website. The download definition draws on IAB Tech Lab's international podcast measurement guidelines.

Why: IAB Australia's own buyer research found that 27% of respondents cited a lack of standardised audience data for digital audio planning. The glossary gives the market a single reference list for an audio ad market worth AU$353 million in FY26, though it records competing definitions rather than settling them, and it predates the changes IAB Tech Lab finalised in September for video podcasts.