How Quantifind Forecast Consumer Intent From Social Analytics.

Home / Companies / How Quantifind Forecast Consumer Intent From Social Analytics.

How do you figure out which lines of social media buzz translate into revenue opportunities? Sentiment analysis is a key part of how Quantifind approach the question by using their own blend of classical stats, probability, natural language processing and machine learning. This means that organisations are able to look at what the broader population are talking about which is relevant to their industry and how it correlates with peaks and troughs in revenue. What Quantifind do is unravel lots of data to expose the indicators which allow effective revenue based modelling of future marketing campaigns.

Quantifind do this by gathering large amounts of structured and unstructured publicly available information such as tweets, comments, blog posts or product reviews and correlate the information with a structured field about activity from the company. The source of the information is dependent upon the vertical but Quantifind say they will always try and pick sources they already have feeds for. The result is a report of sentimental pressure points from the market to see how feedback influences revenue and customer satisfaction.

[vimeo 92737630 w=500 h=281]

Quantifind have products for several different verticals such as the film industry and video game producers where their product is able to better predict who their target audience is for a certain movie or game and also adjust variables to see what would affect customer satisfaction.

Quantifind use datasift to retrieve historical data from Twitter using a custom built feed and use Spark for data analytics. This has led Quantifind to become predominantly a Scala shop after first running Pig for their command line shell which allowed them to try things iteratively and allowed algorithms to be developed collaboratively. The net result here was slow runtimes, difficult to understand error messages and jobs taking an hour to complete but with Spark this could be cut down to seconds shortening job latency and enabling new product features. Complimentary to the evolution of their tech stack, Quantifind have created Sumac which is an open source library for parsing command line arguments in Scala and was developed to define arguments more easily. Quantifind says Sumac gets rid of all the boilerplate, using reflection to figure out the names and types of the argument.

The release of Sumac 0.3 coincides with the company reaching a critical mass where they will be driving the sales operation on the shoulders of an established employee base which includes nine Phd Scientists and solid backing from USVP, Redpoint Ventures, and Andreessen Horowitz.

Related Posts