Watch NOBL's Demo Day presentation from June 21, 2023.
Advertisers prefer high-quality, trustworthy content with highly-engaged audiences — the kind of content and readers that small, local media publishers possess.
Local media needs to attract more advertising than ever as their revenue has been depleted by an online advertising model that favors giant media with huge audiences.
If only there was some way to allow advertisers to target the best small, local media.
NOBL Media, did just that, during the AI & Local Media challenge. NOBL, a startup that uses AI to evaluate and score content, created a list of more than 2,260 high quality local media sites. Lists such as this are commonly used in digital advertising to target ads to a curated set of domains. NOBL’s inclusion list is continuously updated with new sites as they are discovered and is composed exclusively of high-quality local media sites.
The technology evaluates a million pages every day as they flow through programmatic advertising systems and distinguishes high-quality, local media from all the other content with >96% accuracy, precision and recall.
This process starts with an AI model, created during the challenge, which identifies whether the page is from a local media publisher. Next it evaluates the quality of content and page and generates a score using NOBL’s proprietary technology. Finally, the average score of pages for each domain is calculated.
Domains are identified as “NOBL Local Media” sites if at least 60% of the pages are identified as local media and the average score meets the quality threshold score (minimum of 40 pages scored).
The AI model was trained on about 1,200,000 articles that included about 2,100 known local media domains.
The model was evaluated for precision, recall, accuracy, and f1 score on an article-level on a set of approximately 300,000 articles.
The model was then run on a new set of 1 million articles (excluding articles from domains that were already known to be local media) sourced from a live advertising bid stream, during a 24-hour period in June, 2023. This exercise, which simulated our real-time content evaluation process, revealed 193 NOBL local domains that were previously unknown (not in our list of 8000 ‘known’ local media domains). These were television, radio and news sites from smaller communities or small publishers from larger cities across North America.
A partial list of Local Media domains identified by the NOBL Local Media analysis from an active advertising campaign.
These are the types of sites that are easily lost in the increasing cacophony of clickbait, sensational and provocative content on the internet.
Since June, NOBL evaluated 11 million more articles. An additional 193 NOBL Local Media domains have been identified and added to the inclusion list which now has more than 2,450 NOBL Local Media publishers.
Advertisers can target these domains through the use of a private marketplace (PMP) deal ID that NOBL generates. The deal ID directs ads to be placed on pages in this inventory, which is exclusively local media content. Advertisers can be confident that their brand is reaching local audiences and local publishers are rewarded with revenue that enables them to continue serving their communities.
A 2017 survey by the News / Media Alliance by the News/Media Alliance found “41 percent of registered voters believe their local news can be trusted more than the national news.”
NOBL’s Local Media inventory product provides brands with an opportunity to align advertising practices with this trusted source of engagement. By aggregating large numbers of smaller local news outlets into a single set of inventory, NOBL provides the scale necessary for national brands to advertise in local news environments.
Local publishers gain access to larger advertisers, which increases revenue and promises to reverse decades of attrition and atrophy in community journalism.
When both parties benefit, it forges a symbiotic relationship that has natural staying power. This is NOBL’s mission. We are empowering both publishers and advertisers, and, by changing the incentive structure for online publishers, encouraging higher quality news content and a richer and more trustworthy information ecosystem.
Team members: Jacob Bailly, Dr. Elizabeth Earle, Nameer Hirschkind, Cedar Milazzo, Bill Skeet