Gaining profound insights from billions of news reading sessions using data and AI

Andrew Montalenti, co-founder, CTO, Parsley

Gaining profound insights from billions of news reading sessions using data and AI

//
Posted By
/
Comment0
/

By Scott M. Scher, Columbia University

Parsley is a real-time web analytics company that has worked with more than 350 media companies over the last few years. Parsley allows an organization to understand audience attention online with a real time and historic view into what is happening with your content. This helps companies quantify what is really going on when customers visit their websites and provides easy-to-use tools for them to gain insights about customer engagement.

Parsley launched its platform in 2012 and now has 30,000 monthly active users among its 350 media company partners, measuring two million page views per minute, and provides sub-second data latency. This all makes essential data accessible.

Montalenti talked through the inner workings of the Parsley platform. “Parsley under the hood”, uses an in-house built time series database called Mage. At a high level this database takes the best open source distributed computing technology and builds a model around it that allows Mage to take an event from a user engaging with a piece of content and within less than a second create aggregate views summarizing activity about the user’s attention. In order to do this Parsley uses a live streaming system called Apache Storm that has 3600+ real-time processing CPU core, rock solid distributed stream data, and elastic scalability for hourly jobs. Parsley is not just used as a day-to day reporting dashboard for newsrooms, but also to pull multi year reports for commercial teams to see how a campaign has fared.

When Andrew talks about quantifying audience attention, what he really means is that Parsley can create metrics around how users respond to content. They have created a suite of metrics based around; engage time, metadata, page views, return visitors, and publisher data. This provides Parsley the ability to “tell the story behind the story”. By using modern Machine Learning and natural language processing techniques Parsley is able to answer questions that an organization could never answer on their own.

Leave a Reply