Social Media Monitoring Solution

About

The client was a Sunnyvale, California-based tech startup founded by Amazon & Microsoft executives. Due to the explosive growth of social media, tracking what’s being said about a brand or product on the Internet became a huge task. Product companies needed to track a multitude of social networking sites, blogs, message boards, podcasting sites, video blogs, micro blogs and consumer review sites on which conversations were being held. It was imperative to identify major influencers in social media and engage them to get the message out.

 

Big businesses use PR companies and manual systems to identify and track online conversations, but the price is often out of reach to small & medium businesses.

 

The client had an idea for a do-it-yourself (DIY) solution which would enable SMBs to ‘listen, learn and analyze’ across millions of social media sources. The solution had to monitor keywords to discover relevant conversations and provide lists of news sites, blogs and other channels. There was also a need to identify influential writers and bloggers. The client turned to Quantilus as its tech partner to develop the product.

 

Challenge

  • Remote harvest of unstructured data from multiple sources (~millions, 1.3 TB per month).
  • Data transformation (normalize across all sources) analyze data on different parameters i.e. sentiment, timeline, geography, source & demography – at run-time.
  • High-performance distributed data storage and high availability.

Solution

Quantilus was involved from product conceptualization to its complete development and maintenance.

 

The core product service is divided into three sections: Listen, Learn, and Analyze.

  • The Listen section allows search for keywords that users have designated that they would like to monitor (i.e., keep a search out for a recently-released product model to see if they had any complaints).
  • The Learn section offers a listing of news sites, blogs, and Twitter users that have featured content relating to the user’s campaign keywords.
  • Analyze section offers charts and graphs depicting how hot the user’s search terms are on the web using technology developed with Senti-Metrics.

 

We automated the data ingestion process and used big data for distributed storage. We also used NLP to develop Sentiment Analysis for Social Media. Indexing technology was developed to discover, index and catalogue over 100 Million blogs, 250,000 video bloggers, 3.5 million bloggers and 30,000 media properties including dailies, magazines, periodicals & Television / Radio. 167-Node Hadoop cluster was implemented for distributed data storage having ~30TB data. MapReduce was used to map data across sources to produce intelligent analytics at runtime.

 

Other functions include the ability to monitor blogs for new authors, or track which writers have been focusing on a certain topic. The product brings all sources (blog, news, social media) into a single interface that is much easier to manage and to navigate.

 

The product can generate over 120 reports from harvested data such as brand perception, media perception, sentiment analysis, demographics, lexical & text analysis, top engagements / influencers and real-time trends. Results can be sliced, diced, filtered, and sorted to fit the specific needs of a social media program.

 

Meltwater Group, one of the world’s leading global Software-as-a-Service (SaaS) companies, acquired the product for $4 Million within 2 years of idea-to-market. The product is currently serving as the platform to enhance Meltwater’s social media monitoring service, Meltwater Buzz.

eBOOK

Harnessing the Power of

ARTIFICIAL INTELLIGENCE

Need help navigating your AI journey?

Download our eBook for insights and tips to ease your way into adopting AI for your business – the smart and practical way!

Need help navigating your AI journey?

Download our eBook for insights and tips to ease your way into adopting AI for your business – the smart and practical way!