An addressable audience, or simply addressability, refers to the number of individuals a media company has access to that can potentially be reached through an advertising campaign.
In a pre-digital era, addressability was achieved through mailing lists, powered by data provided by Royal Mail in the UK, and the US Postal Service. As email grew in popularity in the 90s, mailing lists took their first step towards being digital, with postal addresses replaced with email addresses.
However, with the explosion of the internet in the late 90s, and rise of social media in the 00s, advertisers became increasingly reliant on programmatic advertising. This created an ‘adtech boom’, with multiple new platforms coming to market that could track and target individuals through the use of third-party cookies, acting as a proxy for identity. These platforms used third-party cookies to create large addressable audiences for brands and media owners alike and for use cases such as re-targeting.
However, these audiences were often ‘unknown’, meaning there was little to no declared data from the individual, such as their email or name, and instead the platforms mainly relied on inferred data such as context or behaviour.
With the abolishment of third-party cookies looming and increased scrutiny from regulators, the advertising industry is increasingly seeking to power their advertising efforts with ‘known audiences’, achieved through a combination of identity and first-party data. This enables advertisers to personalise the messaging and optimise the relevance of their ads to serve the right ad, at the right time, to the right person.
Facebook and Google have dominated the digital advertising space for over a decade, due in part to the extensive addressable audience that advertisers can target at a one-to-one level (a.k.a people-based marketing). While the walled gardens have proven the formula for success, other media owners are now growing their own addressable audience and providing value to advertisers.
Digital advertising, especially where delivered through programmatic methods, requires a media owner or publisher to have an audience that can be targeted through advertising. Meaning the media owner or publisher must have a unique identifier associated with each individual, whether it be an email, cookie, mobile ID or another method. This audience can then be made available to brands to target using the demographic and behavioural attributes that are collected.
For the media owner, providing a large known audience that can be targeted by advertisers to deliver personalised advertising, enables them to provide a viable alternative to the walled gardens of Facebook, Google and Amazon. However, many media owners are just at the start of their journey from an unknown to known audience; implementing paywall and registration layers, encouraged by premium content, newsletters and ability to comment on articles.
The rise of music streaming services such as Spotify, Pandora and Apple Music, as well as the increased popularity of podcasts, have seen new known audiences emerge through audio streaming. Spotify, as an example, requires every user of its service to log in using an email address. This provides them with a unique identifier for every individual using their streaming service.
Spotify’s addressable audience, however, is the over 100 million users who take advantage of the free service, in exchange for being targeted with advertising. This addressable audience can then be targeted by brands using the various demographic and behavioural attributes that Spotify collect in their first-party data.
With over 1 billion Smart TVs and OTT devices in the world, the addressable audiences that are available through these platforms are considerable. While traditional TV advertising has been a mainstay in the marketing mix for some time now, CTV/OTT and other VOD services provide a new opportunity for personalisation and targeting.
As more and more broadcasters create their own streaming services, the vast majority require individuals to sign-up with an email address to access their content. This means these broadcasters are building up an impressive known audience base. This audience can then be made available to brands to target in their marketing campaigns based on the behavioural and demographic data collected by each broadcaster, and for retargeting, using the brands CRM data to target their known audiences directly on CTV/OTT platforms.
In the digital and programmatic space, addressability has previously been achieved through the use of third-party tracking techniques, such as third-party cookies, or more opaque methods such as fingerprinting.
Third-party IDs, such as cookies, have historically been used as a proxy for identity. While this provided tracking and targeting capabilities, the accuracy of these IDs has always been quite unreliable. A common example of this unreliability, is the same individual being treated as several different users, when viewing the same website but switching between devices or browsers. As such, third-party cookies often created a disjointed view of a user leading to irrelevant advertising or content being shown to them.
Additionally, with an increase in regulatory compliance, and browsers moving to block these third-party ID’s, brands, publishers, media agencies and the myriad adtech vendors between, need to replace these soon to be outmoded approaches.
As we’ve covered, third-party cookies have traditionally been used as a proxy for identity, to power an addressable audience. As these IDs can be dropped on both an advertiser and a publisher web property and held in a third-party environment, they allow an individual to be targeted by advertisers across multiple publishers.
While this approach created a common ID to aid addressability between advertisers and publishers, it did so in a way that consumers struggled to understand and offered little control. Regulators have also highlighted the lack of transparency and understanding around third-party cookies to be an issue. In particular the daisy chaining of consent which often occured between platforms sharing third-party ID’s.
The major browsers have all determined in recent months that this approach didn’t protect consumer privacy and therefore started to phase out support for third-party cookies. Apple was the first to make this change, implementing Intelligent Tracking Prevention in their Safari browser in 2017. They were quickly followed by Mozilla and Microsoft in their Firefox and Edge browsers, and then finally Google announced in 2020 they would follow suit, setting a final deadline of 2022.
With the demise of third-party cookies, Universal IDs have been touted as an alternative to tracking and targeting consumers across the web. As these ID’s are delivered in a first-party environment, many have considered them a more future-proof way to collect identity information.
There are essentially two styles of Universal IDs, each with their own challenges. The first is a probabilistic Universal ID. These combine some first-party data with additional user behaviour clues such as the browser used, favourite websites, as well as the screen resolution and download speeds on the accessing device, to stitch together a user profile. This can then, probably, identify that same non-logged-in user across different devices and browsers. This type of approach is commonly referred to as Device Fingerprinting. Fingerprinting is carried out in a non-transparent way, and as such, all three of the major browsers have committed to blocking this.
The second style of Universal ID is deterministic. These work in a similar way to third-party cookies, except the cookie is first-party (meaning it’s placed by the owner of the website). These first-party cookies are then able to track the user’s behaviour on that site. However, they do not provide cross-site functionality as they are tied to the domain of the website owner. To achieve data portability, the IDs and their associated data are shared in some way with the third-party Universal ID provider. The provider is then able to build an identity graph using this data. But this centralisation of data still creates significant privacy and security challenges, as well as resulting in limited visibility into the process for the data subject.
There is an alternative approach which makes the use of deterministic Universal IDs much more privacy-safe, and that is to utilise a decentralised or federated approach to identity, which enables the data to remain with the party that collected the data. Find out more on the InfoSum approach.
The third and final approach to creating an addressable audience is through the collection of PII data, such as email address or mobile ID. As we covered above, streaming services (both audio and video) have demonstrated the ability to create a vast known audience through a simple sign-up requirement. Outside of these two forms of media, we’re seeing other media owners taking similar steps. Many publishers are increasingly putting content behind a sign-up wall. This allows them to associate all behaviour and demographic attributes with that user.
This addressable known audience can then be made available for brands and media agencies to match customers and potential customers against to build an audience segment. However, given the sensitivity of PII data, it would not be acceptable to share this data with a third-party to create connectivity between data sets. Therefore, it is vital that a decentralised or federated approach is taken, find out more in the InfoSum approach.
Connectivity, however, is not the only challenge; the second challenge is scale. We have seen how other forms of media have created significant audiences by putting content behind a login. However, publishers are slightly behind the curve in this approach, and therefore must take action immediately to begin collecting this valuable first-party data.
In the cookie-free world, addressability will become reliant on the move from unknown to known audiences. For media owners, this journey can be achieved through the production of premium content behind a login wall that justifies the user providing their email address to the media owner.
With this unique identifier provided, publishers can begin building a picture of their audience underpinned by behavioural and demographic data. This addressable known audience can then be made available to advertisers to power people-based marketing campaigns.
For advertisers to take full advantage of these new addressable audiences, they too need to build up their first-party data assets. We are already seeing an increased focus on collecting transactional data and tying this data to a user account, and persuading prospective customers to sign-up for newsletters and offers.
However, given the trust the consumer has provided both media owners and brands in handing over their email address, it is absolutely vital they protect this data. The first step in doing so is to not share this valuable data with a third-party.
InfoSum supports addressability by providing a privacy-secure environment to connect first-party data between brands and the media owner. InfoSum takes a federated approach to data management, meaning each party's data remains decentralised in their own controlled environment, known as a Bunker. The owner of the Bunker can then enable another party to conduct analysis statistical against their data to unlock insight and create audience segments. Importantly, this all happens without sharing or moving raw data.
This unique approach enables media owners to make their addressable audience available for brands to analyse and measure overlap between their customers and the known audience data. Audience affinity and potential reach can then be accurately measured.
Importantly, InfoSum utilises cutting-edge differential privacy techniques that ensure no individual in either the brand’s or media owner’s data is ever identified. All results of the analysis are provided at an aggregate statistical level, where various differential privacy techniques have been applied to ensure individuals can not be singled out through analysis.
No results are given or displayed where the number of results is less than the threshold.
Results are rounded, meaning introducing a single individual through granular queries will not cause the result to change
A small level of purposeful inaccuracy is inserted into the results to ensure analysis can’t expose an individual.
Increases the time and cost associated with an attack allowing it to be detected and stopped sooner.
The advertiser can create their ideal audience using the demographic and behavioural attributes in the data sources. This segment can then be flagged for direct activation by the media owner. Because of the patented approach InfoSum takes, neither the publisher or the advertiser have to share their data with the other to achieve activation.