CanCanCan , commonly referred to as Can3 or CanCan for simplicity, is an attention-centric marketing engine powered by data analysis. Can3 studies content, from viral videos to celebrity sightings to blockbusters, as attention units, and investigates the multidimensional relationships between such signals and human incentives. While heavily informed by data models and algorithms, it runs on the synergy of machine and genuine human intelligence.

Can3 is built upon the belief that the future of marketing, or the mechanism that pairs supply and demand, thrives on authenticity, honesty and humanity. The current state of bot infestation, top-down imposition and algorithmic manipulation is a transitional phase.

Ah, why the name? CanCanCan: 勘·瞰·看 (three Chinese characters, all pronounced can, which mean inspection, a bird’s-eye view, and watchfulness, respectively)

Jump to Orientation

Can3 serves CANDYs, CanCanCan Data Yarn files, providing element-level cognition-oriented data on the underlying assets. At the present stage, it mainly focuses on the human behavioral aspect (visual cues, priming effects and salience triggers). In addition to analyzing the innate objects embedded in assets, Can3 also utilizes Brahms, the in-house market intelligence unit, to source compatible products on the external market and curate context-aware merchandise catalogs. An effective marketing strategy by itself, Can3 is part of an initiative to further liberate resources (and consumers) at the object level for a more dynamic market mechanism.

Built as research-first and intended for professional use: Boom•Clap (our native commercial application that helps place products/information into the right attention windows), engagement optimization, affiliate marketing, etc. Some advanced use cases are discussed in the Showcase . In compliance with copyright laws, unless authorized and considered necessary, Can3 does not host content. In some cases, still frames, artistic renderings and social media reposts will be used for reference purposes. If your content is being used in this way and you believe such usage is beyond the scope of fair use, please email us to have it removed.

In Other Projects

CANDY and the variants

At the moment, the system services 4 types of CANDY: Mallow, Licorice, Taffy and Jelly.

While Mallow focuses on TV dramas with strong female leads (Taffy on single characters), Licorice tracks NBA players and pop culture celebrities with distinct styles. In the meanwhile, on a different front, Jelly timely profiles viral trends, TikTok videos, etc. CANDY data is made available for free and on an as-is basis. Redistribution is encouraged. You may download the data packs directly from their respective pages or query the API for better results.

Unless noted otherwise, all data provided by the system shall be considered as mere predictions or projections. Over 90% of the analysis is done by machines.

Can3 gathers external information/intelligence from multiple sources:

• Origins of the resources
• PR companies/agencies
• Social media/online communities
• Internet search engines
• Experts/consultants

Can3 works with Dawn , a computation-based resource manager for creative content, and is under the semi-supervision of the Mill , a computation-based ethics authority. (🌵 Learn more about the ecosystem )