Permutive

Permutive is a data management platform for publishers and advertisers, with cohorts, an identity graph, a clean room and curation, sold by quote.

Permutive helps publishers turn the first-party signals on their own sites and apps into audiences that advertisers can buy. Its Edge technology processes data on the reader's device instead of relying on third-party cookies. That lets a media owner address readers in Safari, Firefox and iOS in real time. Permutive Limited, the London company Joe Root and Tim Spratt started in 2014, builds and sells it. Signals is the data layer. It builds custom cohorts from behaviour and imported data, contextual cohorts from categories, keywords or sentiment, and modeled lookalike cohorts grown from seed users. The Identity Graph links identifiers across domains, devices and sessions, and the Marketplace adds segments from providers such as Eyeota, LiveRamp and TransUnion. Connectivity brings in CRM lists and runs Routing, which exports event data to a cloud warehouse. Clean Room matches a publisher's and an advertiser's audiences without either seeing the other's raw identifiers, while Curation packages publisher signals into audiences sold through SSP marketplaces. The platform is made for publishers whose audience is what they sell, in direct-sold campaigns or curated deals. Advertisers, agencies and retail media teams use it from the other side, bringing their own customer data to match. Ad sales teams answer briefs with Planning, and Insights shows them an audience's engagement, technographics, overlap and the content it reads. Engineers collect events with the JavaScript, iOS and Android SDKs or the direct API. Each contract is priced by quote after a demo, with warehouse routing quoted separately. A typical job starts with a cohort built from reading behaviour, checked for overlap in Insights and sent to Google Ad Manager for a direct deal. The same cohort can reach programmatic buyers through Magnite or Microsoft Monetize, or social campaigns on Facebook and TikTok. An advertiser with hashed CRM lists matches them against publisher audiences in the Clean Room, then models and activates the matched audience. Raw event data streams to BigQuery, Snowflake or Amazon S3 for the publisher's own analysis. Throughout, the site passes each reader's consent to the SDK, and collection and activation follow it.