FAQs

Brand Safety: 

Mantis scans articles to determine a brand safety rating for that content (safe or unsafe for advertising). Brand safety values are generated and associated with an individual URL. These values are passed to the publisher’s ad server as Key Values to target as either a PMP or PG deal. Using Mantis, publishers can unlock more monetizable inventory, as Mantis scans for more outputs than keywords alone resulting in a truer and fairer brand safety classification. 

Mantis also acts as a central library for brand safety settings. These settings can be managed for an individual agency group or advertiser, and is done so fully transparently. 

Advanced Contextual Targeting: 

Mantis allows for the creation of Contextual Targeting segments. Mantis allows for the combination of the Natural Language Understanding outputs (e.g category + sentiment) to reach audiences in the right context, at scale. Using Advanced Contextual targeting means future proofing for a world without third party cookies.

In 2019, Reach developed Mantis to address the ongoing issues across the AdTech industry around blunt keyword blocking and unintelligent brand safety tools. As a publisher, Reach has experienced these issues first hand, and decided to take action by working with IBM Watson to create the most powerful and precise approach to brand safety that is available today.

Mantis scans the page content, excluding any advertising, sidebars or ‘read next’ content. Results are then generated (using the IBM Watson Natural Language Understanding technology) for the following outputs: Category, Concept, Entity, Keyword, Sentiment and Emotion. Mantis gains an understanding of the page content at a semantic level, flagging any unsafe outputs determined by the default settings or the customer’s unique settings. For example, an advertiser may wish to block the concept of ‘death’ . Through analysing the language used as well as the sentiment and emotion of the content, Mantis will determine if the content discusses death and block/unblock accordingly.

There are a number of benefits to choosing Mantis for Brand Safety and Contextual targeting, such as: 

Leverage IBM Watson Natural Language Understanding: 

Mantis utilises IBM’s NLU Machine Learning to provide semantic detail to every article, analysing the meaning of the words and the relationships between them. So that a customer doesn’t have to rely on keyword targeting/blocking alone, Mantis looks at a number of outputs including category, concept and emotion to truly understand the nature and sentiment of the content and provide a true brand safety rating through detailed analysis.

Pre-Published:

Mantis integrates with direct publishers at CMS level, meaning there is no need to implement tags on page, therefore reducing page latency. This also decreases the window of risk between an article going live and a brand safety classification being assigned. With a traditional tag based integration, this window can be anything from one to twelve hours, leaving the content un classified and therefore either un-monetised for long periods of time, or open to advertising when the content is unsafe. Integrating Mantis at CMS level means that the content is classified as safe or unsafe immediately, giving more opportunity for the content to be monetised if it is safe, and giving advertisers the guarantee that they won’t run on unsafe content. 

Customisable: 

Mantis acts as a central library for all brand safety settings, which are fully transparent to publishers. This means brand safety settings across multiple agencies and brands can easily be managed in one place and if an agency or brand makes a change, it will roll out across all publishers using Mantis. Via a managed service, customers can customise brand safety and contextual targeting settings within Mantis.

Control:

Mantis approved inventory is solely available via PMP and PG, giving the customer a much higher level of control compared to running in the OMP. Transacting in this way means that buyers gain access to premium, high quality inventory at scale.

Mantis is currently a managed service, meaning that we will support customers fully from integration through to full account management. Please see our SLA in the ‘Service’ section for further details around service.

Mantis integrates either directly into a publisher’s CMS or via an API for batch processing. Both options require the heavy lifting on our side, although some development is required on the publisher’s side during the integration. Please see Integration with Publisher CMS and Batch integration on our documents site for a deeper technical understanding.

Mantis is charged to the publisher on a cost per URL scanned basis, rather than by number of impressions. If Mantis sees a URL more than once (and the content of the page is unchanged) the previous values will be applied and no charge will occur. Please get in touch to discuss pricing models.

Yes – in its current version Mantis can scan in English, French, German, Italian, Arabic, Korean, Simplified Chinese, Dutch, Japanese and Brazilian Portuguese.

No – Mantis is designed to provide brand safety and contextual targeting only.

In its current version, Mantis scans display inventory across desktop, tablet and mobile web. Our product roadmap includes development for video inventory, as well as AMP and FBIA.