Topsy’s new Pro Analytics service offers powerful insights plumbed from the depths of Twitter’s data stack.
Twitter is the world’s biggest ongoing public conversation. It’s where we turn to praise, blame, celebrate, and joke, and to express how we are feeling, the instant we feel something. Which of course makes it incredibly valuable to marketers seeking to understand what people are saying about their brands online. Today, Topsy is launching a new tool called Topsy Pro Analytics. It’s a powerful measurement system that lets people tap into its massive index of public tweets, going back through two years’ worth of data — a huge dataset, updated in real time, with layers of complex analytics tools. The functionality is breathtaking, for multiple reasons.
“For the first time, we’re making hundreds of billions of tweets available for searching, from seconds ago to years ago,” Jamie de Guerre, Topsy’s VP of products, told Wired.
Yet the service offers far more than simple keyword searching. One of the more interesting things it does is offer related terms. Search on a topic or hashtag, and Topsy will return not only how many times it’s been mentioned and by whom, but what people are most commonly saying about it. It breaks those responses out into further keywords and phrases.
A sentiment-analyses feature tracks how topics trend up and down in terms of positives and negatives over time. It works by looking at every single public tweet sent every day, establishing a baseline sentiment score for that day, and comparing that to individual tweets that mention whatever topic or hashtag or @name you are searching for. It’s the same toolset that powers Twitter’s Political Index. Indeed, according to Guerre, the only difference between that tool and the one Topsy is offering to brands is that the former uses a three-day smoothing effect.
This combination means you can understand why people are talking about certain things.
For example, searching on @RepToddAkin shows a massive decline in sentiment scores, from 50 percent (completely neutral) to 5 percent in a 24-hour period, combined with a massive uptick in the overall activity (people tweeting and using his @name). If you lived in a news void, this might be curious, until you viewed the related topics in a search for his name, and saw the entries for items like “#legitimate rape” “forcible rape” and “redefining rape,” which would inevitably lead you to the news of the day about Rep. Akin’s controversial comments about rape and abortion and the discussion taking place around it.
The system also helps you understand who is saying something. Search for a topic, and it shows you who has had the most influential tweets about it–those that have been retweeted or acted on the most. It also shows what people are linking to on the Web, and how often they’ve linked to it across all services–something that’s often hard to parse in the age of link shorteners like bit.ly or t.co. By analyzing language patterns used by specific users, it can isolate them geographically by country, even if they are not embedding location in their tweets, with what it claims is 90 percent accuracy. So you can see where discussions are taking place.
The real power comes when you begin overlaying one topic with another. This can help you understand why some social media trends happen and others fall flat. Overlay Nike with Adidas and see how the conversation around one brand affects that around another. You can see how hashtags, and language in general, perform on a sentiment level. It’s a truly impressive tool that will certainly help all manner of marketers fine tune their tweets.
It’s also a little unsettling, because this is a classic case of a service where you are the product. Services like Klout simply want to reduce you to a number. Topsy is doing something considerably more complex — you can use it to see how conversations change online, even minute to minute, and which specific tweets and messages drive those changes. It’s a deep dive into big data that has the ability to pluck out not just meaning but also individuals.
Topsy can help corporations and campaigns see which tweets have the most impact on rising or falling sentiment associated with its brand–or the brand of its competitors–and who exactly sent those tweets. It can identify influencers in ways Klout, for example, has never done. Certainly this data is all public already, but Topsy has done more than anyone yet to help us make sense of it. Its new service is both valuable and well intentioned. It will be interesting to see how it’s used.
Topsy has not announced pricing.
No comments:
Post a Comment