Thirty-three terms you will meet when you monetize streaming inventory, defined in plain English by the team that builds the ad server. No marketing language — just what each term means and why it matters to your revenue.
Also: content fingerprinting
ACR is technology built into smart TVs that identifies what is on screen by matching audio or video fingerprints against a reference library. Television manufacturers use it to build viewing datasets, which advertisers then buy for targeting and measurement.
For a publisher, ACR matters because it is one of the few identity signals in connected TV that does not depend on cookies — and because the TV manufacturer, not the publisher, usually controls it.
An ad break is a point in a stream where content playback pauses and advertising plays. A single break contains one or more ad slots, and breaks are typically described by their position: pre-roll before content, mid-roll during it, post-roll after it.
How many breaks a piece of content carries, and how long each one runs, is the publisher's core lever on ad load — the trade-off between revenue per view and viewer tolerance.
Ad decisioning is the real-time process of choosing which advertisement to serve into a given slot. The ad server evaluates every eligible demand source — direct-sold campaigns, programmatic bids, house promotions — and weighs price against targeting rules, delivery pacing and viewer experience constraints.
Decisioning quality is where yield is won or lost. Two ad servers with identical demand can produce materially different revenue depending on how well each one decides. See our CTV ad server for how per-impression decisioning works in practice.
An ad pod is a sequence of ad slots filled together within a single break, requested and assembled as one unit rather than slot by slot. A ninety-second break might be a pod of three thirty-second slots.
Treating the break as a pod rather than as independent slots is what makes rules like competitive separation and frequency capping enforceable across the whole break.
ads.txt and app-ads.txt are public text files in which a publisher lists the companies authorised to sell its inventory. Buyers fetch them to confirm that supply offered as yours really is yours.
Missing or inaccurate files cost real money: many buyers filter out unauthorised supply automatically, so an out-of-date file quietly suppresses demand.
An avail is an available advertising slot in a stream — the atomic unit of connected TV inventory. When a player reaches an ad break, each slot it needs to fill is an avail.
Publishers count avails to size their inventory, and measure what share of them earned revenue as fill rate.
Beaconing is the firing of tracking calls at defined playback milestones — start, first quartile, midpoint, third quartile, complete — so buyers can verify an advertisement was actually delivered and watched.
In server-side ad insertion the server fires these beacons on the player's behalf, which is more reliable than client-side tracking but requires the server to model playback accurately.
Competitive separation is a pod-building rule that prevents advertisers from the same category appearing in the same ad break — two car brands back to back, for example.
Direct advertisers frequently require it contractually, so an ad server that cannot enforce it across a whole pod limits what a publisher can sell.
CPM means cost per mille — the price paid for one thousand advertising impressions. It is the standard pricing unit across connected TV, where CPMs sit well above most other digital video because of the screen and the audience.
CPM alone does not describe revenue. A high CPM on poorly filled inventory can earn less than a lower CPM at full fill, which is why yield is the better measure.
CSAI is ad insertion performed by the video player on the viewer's device. The player requests advertising separately from content and plays it through its own ad module.
CSAI is simpler to deploy and gives richer client-side signals, but it introduces buffering at every break, depends on player SDKs, and is visible to ad blockers. Compare SSAI.
CTV refers to any television set connected to the internet and used to stream video — whether through a smart TV platform such as Samsung Tizen or LG webOS, or an external device such as Roku, Amazon Fire TV or Apple TV.
CTV is distinguished from broader OTT by the screen: the living-room television, with the shared viewing and premium pricing that come with it.
DAI is the practice of replacing advertising in a stream on a per-viewer basis at playback time, rather than baking fixed advertising into the content file. Two viewers watching the same programme see different ads.
DAI is what makes streaming inventory addressable at all, and it applies to both live and on-demand content. It is usually implemented through server-side ad insertion.
A data clean room is a controlled environment where two parties match audience data for targeting or measurement without either side exporting the other's raw records. Only aggregated results leave the room.
Clean rooms became central to CTV as third-party identifiers declined, because they let a publisher and an advertiser collaborate on first-party data without either handing over its asset.
A deal ID is an identifier that ties a programmatic transaction to negotiated terms — an agreed price, a specific audience, a defined slice of inventory. A private marketplace, or PMP, is an invitation-only auction run using these deals.
Deal IDs let a publisher keep the efficiency of programmatic buying while preserving the pricing control of a direct sale.
A DSP is the system advertisers and agencies use to buy programmatic inventory, bidding into auctions on their behalf according to budget, targeting and performance goals.
A DSP sits opposite an SSP: one represents the buyer, the other the seller.
FAST stands for Free Ad-Supported Streaming Television — a linear, scheduled channel delivered over the internet and funded entirely by advertising, with no subscription.
FAST channels behave like broadcast television in scheduling but like digital in targeting, which makes ad break management and pod filling the central monetization problem for their operators.
Fill rate is the share of requested ad slots that were actually filled with a paid advertisement. If a player requests ten thousand avails and nine thousand return an ad, fill rate is ninety per cent.
Unfilled avails are lost revenue and, worse, a visible gap or slate for the viewer. Low fill usually points to restrictive floors, thin demand, or a timeout set too tight.
Frequency capping limits how many times a single viewer sees the same advertisement within a defined period — for example, no more than three exposures per day.
Capping is a viewer-experience control with direct revenue consequences: too loose and viewers abandon, too tight and inventory goes unfilled. In CTV it is hard to enforce because household identity is fragmented across devices.
Header bidding solicits bids from multiple demand sources simultaneously before calling the ad server, so every buyer competes on the same impression at the same moment.
It emerged to replace the sequential waterfall, and in streaming it is usually implemented server-side, often via Prebid Server.
Manifest manipulation is the rewriting of a streaming playlist — the HLS or DASH manifest — so that advertising segments are listed inline with content segments. The player simply plays the list it is given.
This is the mechanism behind server-side ad insertion: because ads and content are indistinguishable in the manifest, playback is seamless and ad blocking is impractical.
OpenRTB is the IAB specification defining how programmatic bid requests and responses are structured between supply and demand platforms. Version 2.6 added connected TV features including podded requests and richer content signalling.
OpenRTB support is what lets a publisher's ad server talk to the wider programmatic market without a bespoke integration per buyer.
OTT describes video delivered directly over the internet, bypassing traditional cable and satellite distribution. It covers every device — televisions, phones, tablets, browsers.
OTT is the broader category; CTV is the television-screen subset of it. The distinction matters commercially, because CTV inventory prices higher.
Pod bidding auctions every slot in an ad break together, letting the ad server optimise the whole pod — sequencing, separation, total value — rather than deciding each slot in isolation.
Slot-by-slot decisioning leaves money on the table because it cannot trade a weaker bid in one position for a stronger combination across the break.
Prebid Server is an open-source, server-side implementation of header bidding that collects bids from multiple demand partners in a single call, rather than in the viewer's device.
For streaming it is the practical option: connected TV devices have limited processing headroom, and moving the auction server-side removes that constraint and cuts latency.
SSAI stitches advertising into the video stream on the server, so that ads and content arrive at the player as one continuous stream. The player does not know where content ends and an advertisement begins.
SSAI is the dominant approach in connected TV for three reasons: it eliminates the buffering that client-side insertion causes at every break, it needs no ad SDK on the device, and because ads are indistinguishable from content in the manifest, ad blockers cannot remove them. The trade-off is that the server must handle beaconing itself. Our ad server performs SSAI at broadcast grade.
An SSP is the system publishers use to offer inventory to programmatic buyers and run auctions for it, connecting one seller to many DSPs.
An SSP is not the same thing as an ad server. The ad server owns the final decision across all demand — direct, programmatic and house; an SSP supplies one category of that demand.
A timeout is the maximum time an ad server waits for bids before making its decision. Wait longer and more demand can respond; wait too long and the viewer sees a delay or a blank slate.
In connected TV the timeout budget is unforgiving, because any delay lands on a television screen where viewers are least tolerant of it. This is why decisioning latency is a revenue metric, not just an engineering one.
VAST — Video Ad Serving Template — is the IAB XML format an ad server returns to tell a player which creative to play, how long it runs, and which tracking events to fire.
VAST is the lingua franca of video advertising. Any ad server that speaks it can integrate with almost any player without custom work, which is why standards support is the first question in a migration.
Viewability measures whether an advertisement had the opportunity to be seen, against thresholds for how much of it was on screen and for how long.
In connected TV viewability is usually high by nature — video plays full screen — so buyer scrutiny shifts to whether the impression was genuine at all, making invalid traffic detection the more contested measure.
VMAP — Video Multiple Ad Playlist — is an IAB format that describes where ad breaks sit within a piece of content and what should fill each one. Where VAST describes a single advertisement, VMAP describes the schedule of breaks around it.
VMAP is how a publisher expresses its ad load to the ad server declaratively, rather than hard-coding break positions into each player.
VOD is content a viewer starts whenever they choose, as opposed to a scheduled linear stream. It may be ad-supported (AVOD), subscription-funded (SVOD), or transactional (TVOD).
VOD is easier to monetize than live streaming because break positions are known ahead of time and inventory can be forecast, but it lacks live's concurrency spikes and premium pricing.
A waterfall is a legacy monetization method that offers inventory to demand sources one at a time in a fixed order, passing down the chain until someone fills the slot.
Waterfalls systematically leave revenue unrealised, because a buyer further down the order may have paid more than the one who filled it. They also add latency at every step. Unified auctions and header bidding exist to replace them.
Yield is the revenue a publisher earns from its available inventory, usually tracked as revenue per thousand available impressions rather than per thousand delivered ones.
Yield is the metric that matters because it captures price and fill together. Optimising CPM alone can reduce yield by pricing inventory out of the market; optimising fill alone can reduce it by selling too cheaply.
Something missing, or a definition you would put differently? Tell us at contact@gogoctv.com — we maintain this page.
Book a 30-minute technical walkthrough of the ad server — decisioning, server-side insertion and reporting, live on your streaming inventory.