June 1, 2021

Don’t chase cookies, learn how CTV targeting really works – The State of CTV Targeting (1/2)

ID headwinds & turbulence, but not for CTV.

The programmatic industry for the last 10+ years has been underpinned by 3rd party cookies & mobile device IDs. However, global privacy regulations & updates from browsers & mobile operating systems have forced digital advertisers, their agencies & DSPs to start exploring new & alternative ways to target their audiences.

These audiences have been migrating from traditional TV, digital & social channels to free ad supported streaming services in record numbers over the last several years and couple this trend with all of the identity headwinds faced in other channels, marketers are now starting to increasingly turn their focus towards CTV as a way to connect with their audiences. 

CTV operates in a very different environment than desktop browsers & mobile app stores and the identifiers passed today are not yet subject to the same headwinds as those in display advertising. This chart was recently published by the IAB & provides a good overview of what is actually happening in the desktop display & in-app advertising ecosystem when it comes to identity. The main identifiers passed in CTV advertising today by the big streaming apps are deterministic RIDA (Roku), TIFA (Samsung), AFAI (Amazon), VIDA (Vizio) identifiers. This means advertisers, their agencies and DSPs can still find scaled audiences in streaming environments where they are now increasingly struggling to do so in desktop display & mobile in-app. 

In this recent episode of the Publica CTV Explainer series, ID5 Co-Founder & CEO Mathieu Roche explains how streaming publishers are currently building out their first party data strategies & why a robust & privacy compliant identity offering will drive increased demand for streaming PMP deals this year. 

  • CTV publishers are in a really strong position when it comes to 1st party data collection.
  • More programmatic advertising budgets are going to be directed towards CTV publishers, however, to capture it you need a robust identity strategy. 
  • CTV publishers are uniquely placed to build 1-1 conversations with their viewers and this will position them well when navigation global privacy regulations.

Although IP is a targeting feature that is widely used today in CTV ad buys, many say this may be obfuscated due to privacy concerns and the potential PII nature of IP. However, despite the restrictions unfolding in the desktop & mobile sector, there has been no sign yet that companies such as Roku, Samsung, Amazon or Vizo will restrict the use of IP in CTV advertising.   

Today, an advertiser has 2 ways to buy streaming inventory:

Direct deal: from a publisher’s sales team directly via an insertion order at a fixed price. This graphic from the IAB below illustrates the flow of a direct order fulfilled via a streaming publisher in 2021.

Programmatic buy: using a DSP that targets CTV inventory on Ad Exchanges. This graphic from the IAB illustrates how publishers are increasingly selling their streaming inventory in 2021. Streaming publishers are quickly embracing programmatic methods of ad delivery, as greater demand & increased yield are tied with this practice and enables automation with the buy side. PMP is the most used transacting mode for CTV in 2021.

The basic underlying technology supporting CTV targeting

Within the CTV environment – the ad is being stitched in real-time within the video content delivered to the audience over the internet – which implies communication between the ad-server and the ad-stitcher operating behind the scenes when an ad-break occurs.


To allow for the match between desired targeting and upcoming ad opportunity, it is required that some type of information or meta-data is passed from one server to the server: it could be in the shape of a deal-id when transacting directly, or contained within any parameters of a bid request passed to an SSP when transacting programmatically. No matter how the signals are getting to the advertisers, there are today a wide range of targeting layers available to marketers when it comes to buying CTV today.

1) Contextual targeting – the ability to target where the ad plays 

A – CTV offers basic content-related targeting, in parity with TV
  • Inventory: you can set and target per channel.
  • Content: targeting based-off any meta-data related to the program, for example content title, show or distribution network.

B – CTV embarks digital targeting capabilities similar to the web, mobile or social environments

This layer of targeting available to marketers rely on the information provided by the app playing the content on the TV set: Samsung TV+, CBS, Philo, etc.

  • Privacy: ability to target/untarget users based on privacy settings
  • Geo: country, city, zip code, DMA
  • App: app OS (Samsung TV, FireTV, Roku, etc.)
  • Stream: Live with VOD, DVR (recorded livestream)

C – Advanced targeting based on content/context
Advanced targeting at the ad-pod / session level

When the publisher is equipped with state of the art ad-podding technology like Publica, it can offer adbreak or session exclusivity.  Advertisers have the possibility to secure the position of the ad in the adbreak for a premium fee, or the option of “buying out” the ad-break to be the single ad within a category – for instance, paying higher CPM to be the only automotive ad within an ad-break or even within a viewer session. 

Advanced contextual targeting utilizing 3rd party technologies

In this scenario, the advertiser is utilizing technology relying on predictive intelligence that enables them to build segments of audiences with the use of contextual data, potentially enriched with 3rd party data. The simplest example would be the “food -interest” segment displayed below. To effectively target this segment, the publisher has to pass down some type of proprietary ID provided by the service that is later used at the DSP/SSP level to identify opportunities matching segments when an ad break occurs.

Image source: IRIS.TV

To better understand the role of context in CTV advertising, check out this Publica Q&A with the Co-Founder & COO of IRIS.TV, one of the leaders in the contextual advertising space.

D- How does ACR technology work?

ACR tech has the potential to capture all types of TV viewing: linear, video on demand, OTT, commercials and video games. That’s made it especially attractive as overall viewership splinters.

ACR identifies content by matching snippets of content against a reference library of nearly everything that’s on TV. The technology must be able to make a correct match, and the content must be in the reference library in order to be matched.

For example, at Samsung Ads, the approach is to leverage ACR data from more than 50 million smart TVs. This ties a thread from linear to digital viewing, allowing advertisers to manage video campaigns across apps, bundles and channels in much the same way that a consumer does. 

More here in these AdExchange & Digiday articles.

2) Audience targeting – the ability to target who is watching the ad

Audience targeting utilizes data segments piped in the selling/buying process described earlier. There are 3 main types of data sets available today in the CTV world:

A) First-party Data
Publisher First Party Data 

Most likely in the shape of a list of Device IDs, it is uploaded via the ad-server, it can be leveraged and targeted by any direct or programmatic demand who wishes to leverage any publisher proprietary data, such as watching habits. CTV publishers would most likely collaborate with advertisers in this scenario to help them reach specific campaign goals: narrow down targeted audience, run a forecast and buy against it, using a deal ID.

Advertiser First Party Data 

Many advertisers collect data about consumers (in compliance with geo specific privacy regulations) and leverage this to place tailored & targeted CTV ad buys. Outcome? It enables advertisers to share their proprietary data with publishers and run campaigns against it. For instance, a car brand seeking to target an email list of subscribers can share this data set with LiveRamp, who will operate the matching of emails with TV identifiers in a privacy compliant way and then run it against the publisher’s inventory monetized via Publica.

B) Third Party Data 

Similar to contextual 3rd party data providers, CTV has leaders in 3rd party audience management, in the name of LiveRamp, TruOptik, Cadent et al. Once integrated with the ad-server, it enables publishers to access prebuilt audience segments from hundreds of anonymized data profiles and achieve fine segmentation of their audience to eventually pass signals down to the advertiser in real-time when users from any segments turn the TV on. 

The identifiers used to match TV sets to individual users within a household are Device ID and IP, enabling identity graphs and predictive technologies to kick in and unveil user profiles within the same household.

Here is a Guide on audience based targeting in CTV

From an advertiser standpoint, the buying process gets a lot easier especially from a lower funnel standpoint. The current trends in the market show great traction with new buy-side tools enabling marketers to reach targeted audiences in CTV directly from their DSPs offering access to data from companies such as LiveRamp et al.

Here is a video of LiveRamp’s CSO Jay Prasad talking about CTV data infrastructure & how it can improve a publisher’s advertising yields. 


Advertisers are increasingly targeting audiences on ad funded streaming services & the publishers who have a robust identity & data offering are going to be well placed to increase the share of ad budgets they capture from the publishers that are yet to activate their data strategies. 

Programmatic advertisers are looking to target streaming audiences via data enriched private marketplaces and companies like Publica now offer the tools & technology that enable publishers to build the advanced ad breaks that CTV Apps & smart TV manufacturers need to be running to ensure they capture demand from advertisers & their DSPs.

CTV publishers are well placed to grow their share of traditional TV, digital display & social ad budgets this year, but only if they put audience data & advanced ad break controls at the centre of their monetization strategies. 

In part 2 we will focus on CTV trends, what it means to be ‘open” Vs “walled’ & the growing importance of viewer ‘consent’. If you would like to better understand any of the areas we have covered above, please contact Publica here and we will be happy to help.

Posted BY: