Topsy Uses Twitter Data To Analyze Search, Tweets
Topsy wants to become a Google predictive analytics tool, but rather than using search data to drive campaigns, the tool will use Twitter data. Social signals and conversations will increasingly play a more important role in media buying and the way marketers predict demand for products and services.
Mining the tweets will help track consumer sentiment across platforms. It now indexes every message since the first tweet posted in 2006 -- about 425 billion pieces of content from photos to linked pages. The initial data archive dated back to 2010. The platform analyzes the amount of times Twitter users cite a specific piece of content or topic.
Topsy now stores more than 425 billion tweets, videos, images and blog posts, which the company estimates as more social data than either Bing or Google. Aside from mention counts, exposure and comparative analyses, the tool enables tracking of sentiment around any topic, hashtag, group of keywords, or Twitter handle. It infers location for more than 95% of all tweets, allowing for search and analysis by geographic location.
A marketer or a financial analyst using Topsy's analytical tool can predict the popularity of the Apple device several days before it is released, per Topsy SVP Jamie de Guerre. Even if the device receives negative media from the press in online publications, the sentiment from Twitter tweets could show a positive response from consumers for the device.
That prediction becomes valuable for several businesses and audience-buying. A hedge-fund manager or investor could make a decision to buy stock in Apple prior to the release of financial information based on the sentiment of current and historical tweets. An electronics supply chain manager might forecast additional demand based on tweets, or AT&T and Verizon might want to purchase a few more iPhones to support consumer requests for the handset.
Search engine marketers could optimize and benchmark campaigns against prior trends. "If we search for 'Food Network' and want to understand what people have talked about during the last 30 days, we can identify the celebrity chiefs, keywords and hashtags that come up," he said. "It becomes valuable for use in search and social media campaigns."