OTT advertising is delivered inside streaming video that reaches viewers over the internet, on smart TVs, streaming boxes, phones, laptops and game consoles, instead of traditional way through cable or satellite. OTT ads can be targeted to a household or device, bought directly or programmatically, and measured impression by impression.
Streaming is now where most TV ad money is heading. In 2026, digital video is expected to take 61% of US TV and video ad spend, with linear TV at 39%, according to the IAB’s 2026 video ad spend report. Most OTT guides are written for the buyer. This one covers both sides: how advertisers buy OTT, and how publishers price, deliver and protect it.
What Is OTT Advertising?
OTT stands for “over-the-top.” The OTT advertising is any ad content goes over the top of traditional distribution, straight from the streaming service to the viewer through the internet. That covers a wide range of viewing. A household watching an ad-supported tier on a smart TV, a commuter watching on a phone, and a sports fan watching a free FAST channel are all OTT audiences.
One distinction matters before going further. OTT describes how the content is delivered. CTV (connected TV) describes the screen it is watched on. All CTV is OTT, but not all OTT is CTV. Our guide to CTV vs OTT vs linear TV covers the difference in full.
How OTT Advertising Works, Step by Step
- A viewer starts a stream: They open a streaming app on a TV, phone or laptop.
- The stream reaches an ad break: The app sends an ad request that describes the content, the device and, where permitted, signals about the household.
- The ad server checks committed deals first: If a direct campaign needs that impression to stay on schedule, it takes the slot.
- Remaining impressions go to auction: Buyers bid in real time through programmatic channels.
- The winning ad plays: It is either stitched into the stream on the server (server-side ad insertion) or played by the app’s video player.
- Delivery is recorded: The impression, whether the ad played to the end, and the device all feed reporting and billing.
The whole sequence happens in a fraction of a second, every ad break, for every viewer.
Types of OTT Ads and Where They Run
Common OTT ad formats:
- Pre-roll: plays before the content starts.
- Mid-roll: plays during natural breaks, much like a TV commercial.
- Post-roll: plays after the content ends.
- Pause ads: appear when a viewer pauses the stream.
- Interactive and shoppable ads: let viewers respond, scan a code or buy from the screen.
Where OTT ads run:
| Platform model | How viewers pay | Advertising | Examples |
| AVOD (ad-supported video on demand) | Free | Ads fund the service | YouTube, free on-demand apps |
| Hybrid subscription | Lower-priced plan with ads | Ad tiers alongside ad-free plans | Ad-supported tiers of major streaming services |
| FAST (free ad-supported streaming TV) | Free | Linear-style channels funded by ads | Pluto TV, Tubi, The Roku Channel |
| vMVPD (virtual pay TV) | Subscription | Live channels with ad breaks | YouTube TV, Sling TV |
Most large media companies now run several of these at once. That is good for reach, and it is also where many revenue problems start, as covered below.
How OTT Audience Targeting Works
OTT ads can be targeted far more precisely than a traditional TV spot. Common options include:
- Household and device: reach specific households or screen types.
- Demographic: age, income or household make-up, where data allows.
- Behavioral: viewing habits and interests.
- Contextual: the genre, show or live event being watched.
- First-party data: the publisher’s own subscriber and viewing data.
First-party data is the publisher’s strongest asset, because advertisers cannot buy it anywhere else. To use it safely, advertisers increasingly want to match it against their own customer lists without either side handing over raw records. That matching happens in a secure environment such as a clean room. Platforms built for privacy-safe audience collaboration make this practical, and strong data governance for first-party audience data keeps it compliant.
How OTT Advertising Is Bought
There are four main routes:
| Route | How it works | Best for |
| Direct (guaranteed) | Advertiser and publisher agree volume and price up front, often during the upfronts | Big brand campaigns and premium content |
| Programmatic guaranteed | Same commitment, executed through programmatic systems | Guaranteed delivery with automated execution |
| Private marketplace | Invitation-only auction for selected buyers | Premium inventory with some price discovery |
| Open auction | Real-time bidding open to many buyers | Scale and flexible budgets |
Most publishers sell through all four. The hard part, covered below, is deciding which route should win each impression.
What Drives the Price of OTT Advertising
OTT is usually priced per thousand impressions (CPM). Some campaigns are priced per completed view (CPCV), so the advertiser pays only when the ad plays to the end.
The price of any single impression depends on several factors:
- Screen: ads on a TV screen usually command more than ads on a phone or laptop.
- Content: live sport and premium series attract higher prices than general on-demand content.
- Targeting depth: the more precise the audience, the higher the price.
- Completion: non-skippable placements with high completion rates are worth more.
- Timing: demand peaks during the upfronts, the holiday quarter and major live events.
- Buying route: guaranteed deals trade price for certainty; auctions trade certainty for price discovery.
For publishers, the risk is pricing all of this as one blended “TV” or “streaming” rate. Premium screens and premium moments end up sold for an average price.
How OTT Advertising Is Measured
OTT measurement is far more detailed than traditional TV ratings. The core metrics are:
- Impressions: how many times the ad was served.
- Completion rate: how many ads played to the end. Our guide to video completion rate explains how it is calculated and benchmarked.
- Reach and frequency: how many households or devices saw the ad, and how often.
- Attribution: what happened after the ad, such as site visits or purchases.
The challenge is that each platform and currency counts differently. Publishers that bring these numbers into one measurement view across linear, streaming and digital can show advertisers real performance instead of an average. On the viewing side, OTT analytics shows how audiences actually engage with the content around the ads.
Where OTT Revenue Leaks, and How Publishers Fix It
For advertisers, OTT is about reach and precision. For publishers, the bigger question is how much of that value they actually keep. Most revenue leaks happen between systems, not inside any single one.
Leak 1: The same viewer is counted and paid for twice
A large advertiser wants to buy a publisher’s whole portfolio in one go: linear, streaming and FAST. But each platform is planned and sold separately. The same viewer gets counted on each, the advertiser pays more than once for one person, and the publisher’s own teams end up competing for the same audience. The advertiser sees wasted spend. The publisher closes a smaller deal than its portfolio deserves.
This is the problem of converged TV, and it is solved less by buying new software than by joining four things up:
- One inventory and forecast view. Planners see and reserve availability across linear, streaming and FAST in one place instead of three. Inventory forecasting across platforms makes that possible.
- One order. A single insertion order spans every platform, which is what makes an audience-guaranteed, cross-platform buy possible.
- Cross-screen frequency control. The direct fix for paying twice for one viewer. It only works when households and their devices are tied together, the same identity foundation used for first-party targeting. Real-time campaign monitoring keeps frequency in check while the campaign runs.
- Single yield governance. One team prices guaranteed and auction demand together, from one revenue view across linear and digital, so two sales channels stop competing for the same impression.
All four depend on a shared data foundation. Modernizing the media data platform is usually the first step.
Leak 2: Nobody knows what a guaranteed deal cost
Every OTT ad slot can be filled by a committed direct deal or by the auction. When the committed deal takes the slot, most publishers never record what the auction would have paid. So nobody knows what that choice cost.
This matters most in live sport, where auction prices spike. Our look at protecting revenue during the biggest live events covers why.
The answer is not “always take the higher bid.” Committed deals must be delivered, or the publisher owes make-goods. The answer is being able to price the choice: log the bid you passed on, compare it with the make-good risk, and decide with both numbers in front of you.
Leak 3: Delivery problems are found too late
Campaigns fall behind schedule, drift to the wrong audience or show too often to the same viewer. Too often the publisher finds out when the advertiser calls, after the money is spent. Our guide to where CTV ad delivery actually breaks covers the most common failure points.
OTT Advertising Trends to Watch in 2027
Six shifts will shape how OTT advertising is bought and sold in 2027.
- Streaming closes in on TV’s ad budget
US CTV ad spend is forecast to reach about $42 billion in 2027 (Adwave, citing eMarketer), while linear TV ad spend is expected to fall to $56.83 billion (eMarketer). eMarketer expects CTV to overtake traditional TV in 2028 (AI Digital, citing eMarketer). If those forecasts hold, 2027 is linear’s last year in front. Publishers that still price streaming as an add-on will feel the gap first.
- AI agents start buying and selling OTT inventory
Two-thirds of US ad buyers plan to focus more on agentic buying, according to an IAB survey (eMarketer). Agent-to-agent CTV deals are already live using the Ad Context Protocol, with a person on the publisher side approving each deal before the ads are scheduled. For publishers, the 2027 question is whether buyer agents can find and understand their inventory. Clear, structured information about inventory, audiences and formats becomes a sales asset, and AI that prepares ad sellers for every buyer conversation helps teams keep up.
- More ad tiers, more supply, more pressure on price
Ad-supported tiers keep expanding. When Amazon Prime Video moved its users onto an ad tier by default, streaming ad supply jumped and prices fell (eMarketer). More supply gives buyers more choice. Publishers protect yield by pricing premium screens and premium moments separately, and by forecasting inventory accurately.
- Retail media moves onto the TV screen
One in five CTV ad dollars is projected to go to retail media by 2027, according to eMarketer (AI Digital). Retailers bring purchase data. Publishers bring audiences and premium content. The two meet in clean rooms, which turns privacy-safe audience matching into a revenue capability.
- Buying linear and streaming as one
Planning tools now split budgets across streaming and linear TV together, market by market (Cynopsis). More advertisers will expect publishers to sell the same way: one plan, one order, one frequency cap across screens. That is the converged TV problem described in Leak 1 above.
- Programmatic keeps growing, and so does the cost of guessing
More than 84% of CTV is already bought programmatically (StackAdapt). As auction demand grows, so does the value of knowing what each guaranteed deal displaced (Leak 2 above). Publishers that log the bids they pass on can price their direct deals with evidence.
Is OTT Advertising Worth It?
For advertisers, yes, for most campaigns. OTT reaches audiences that no longer watch scheduled TV, with targeting and measurement that linear cannot match.
For publishers, the question is not whether to sell OTT but how. Linear TV is shrinking: in 2025, a year without the Olympics or national elections, US linear TV ad spend fell 14.4%, according to the IAB. The publishers that win are the ones that price streaming on its own value, sell the portfolio as one, and keep the revenue that would otherwise leak between systems.


