The IAB’s Content Taxonomy enables a consistent, easy-to-understand language across the entire advertising ecosystem (advertisers, publishers and platforms) to describe content in order to make decisions on relevance and appropriateness of ads. By speaking the same language, we can promote more efficient & effective buyer-seller relationships whilst maintaining great user experiences.
Carbon’s taxonomy was originally geared towards the purchase funnel due to an early focus on retail media publishing. Having moved beyond retail media we’ve been adapting our taxonomy to include a much wider range of keywords and taxonomy nodes, whilst maintaining some of our unique taxonomy nodes such as brands. This led us to adopting the IAB’s Content Taxonomy 2.2.
Adopting the IAB categorization is key for market fit – meaning easier integrations with clients, partners and advertisers that use it as their standard. For example we can use it with GAM when using the categories for contextual targeting; whilst it makes for easier discussions between our clients and their prospective direct advertisers when exploring audience targeting. Furthermore, it makes for a more consistent experience across insights, audience segmentation and activation.
Going further than the IAB
As a tech platform that had privacy built-in from the start, we’re very conscientious as to what should be allowed in our taxonomy; disabling the parts for what can be extremely sensitive information, such as religion, and medical conditions. We’re also a tech platform with some well ingrained USPs of which our automated categorization is one; so using the IAB taxonomy provides a foundation, but we want more. Carbon goes above & beyond by continuing to exclude sensitive signals, whilst including custom signals such as brands and keywords.
Our automated categorization extracts meaningful content of an article, which is identified and labelled based on the IAB Tech Lab Content Taxonomy 2.2; as well as our custom nodes, brand nodes and keywords based on the specific needs of each client. The resulting categorisation of an article’s context can be used for selecting the most relevant ads for users.
To do this, Carbon combines NLP (Natural Language Processing) and on-page knowledge to build word vectors that reveal relationships between words that can then be more accurately mapped to the most relevant taxonomy nodes. Our proprietary algorithms then score data to determine the strength and relevancy of keywords. For advertisers this means improved brand relevance and contextual targeting to precisely align their ads with relevant content and drive better ROI. For publishers it means the ability to retain the value of their ad inventory while providing a better user experience.
How this will impact activation
Adopting the IAB content taxonomy can enable everyone to speak the same language; making PMPs and direct sales easier to negotiate, whilst powering more scalable solutions such as contextual advertising.
For direct deals, for instance, it gives publishers the ability to develop their own audience segments on their own properties, with universally-defined attributes that advertisers can recognise. Similarly they can offer more granular & specific audiences on the open market based on common attributes. For contextual targeting to work, companies must adopt these standards to give medium-to-large publishers – with less 1st party data – the opportunity to monetise consentless and/or unaddressable consumers.
Beyond activation, with the shifting nature of identity, common taxonomies will become increasingly important for insights, measurement and attribution too. For instance, clean rooms are already growing in usage as a way for publishers and advertisers to pool their data anonymously but with common language to determine key insights such as cross-over.
How will this impact current Carbon users?
All audiences that have already been created using Carbon’s old taxonomy will still function without any additional setup. The old intent categories have been mapped to the new categories under the IAB content taxonomy. We’re also keeping any intent categories from the old taxonomy that have significant volume. It is worth noting that there are some categories that are now listed under “deprecated” and will eventually be removed from the system. All of these had low volume and are not likely to impact the reach of any current audiences significantly. Deprecated intent nodes will be removed automatically from audience attributes in most cases, if you happen to see one in one of your audiences we recommend removing it manually.
The attribute field for intent in the modal for audience creation now displays the IAB content taxonomy instead of the old Carbon taxonomy, so users are now able to take advantage of the new categories.