A Data Science Central Community
You’ve likely been experiencing a deluge of online information coming at you in recent years — an overwhelming number of status updates, e-mails, tagged images and so forth. You’ve probably also seen, and potentially been alarmed by, the growing accuracy of targeted advertisements — “People You May Know,” and other “offers” online.
As the quantity of irrelevant information has exploded online, so too has the market for the delivery of targeted offers and information. Social networks, in theory and in practice, expose many people to contact and influence. Without precise models, people will continue to be bombarded with ineffective offers and other irrelevant information. Predictive analytics, a branch of data mining concerned with predicting future probabilities and trends, applies a filter to users’ online interactions with the aim of delivering more value from a sea of irrelevance.
With increased value comes the potential for social networks to make money as well. Here’s a look at some specific ways in which predictive analytics will make social networks money.
Many recruiting sites out there on the web, from LinkedIn to SelectMinds to Monster, promise to be able to match candidates with job requirements in unique and increasingly accurate ways. Predictive analytics is at the core of their business model, as it automates the process of making these matches.
When a recruiter posts a job description, a predictive algorithm runs through candidates and calculates compatibility. The technology is, in many cases, embedded in search applications. The most accurate and efficient of these analytics will deliver the most value and see the greatest adoption over time. Those recruiting and talent acquisition sites that allow businesses to leverage the existing social networks of their current and former employees are the best positioned to monetize their users’ employment data in new ways. Businesses can get value from these existing networks without the time and resource commitment it takes to build their own.
As sites like Twitter and Facebook gain value to the business world, many companies have cropped up to analyze and establish what the sentiment is of the collective online intelligence and also to identify individuals with influence and authority. Companies including Klout, ViralHeat and Radian6 all scan blogs and other social media channels with predictive models to determine if the content surrounding a brand or person is negative, positive or neutral. As this information becomes increasingly valuable to businesses of all sizes, these sentiment analysis companies are expected to grow rapidly.
Social media channels are open to everyone. Day traders, retail investors and analysts are cruising around on Twitter and Facebook. What these types of people say and do online is not insignificant in an era when [Flash Crashes and Fat Fingers] are being closely scrutinized and regulated. New models are cropping up to predict stock fluctuations based on Twitter posts. Similar to sentiment analysis, these companies are able to look at the total number of tweets, as well as positive and negative comments to predict whether a stock price will go up or down. These types of companies will become a hot commodity as investors begin to rely on the wisdom of crowds.
No one likes to be bombarded with irrelevant offers and content while using their favorite social network. But the more active you are online, the more effectively predictive analytics can work to deliver targeted and relevant offers.
Sometimes it feels like Facebook knows you better than you know yourself. RSVPed “Yes” to that big gala? You may see a discount offer for Saks. [Are you a woman between the ages of 18 and 34? A Facebook ad may tell you how you can lose those extra inches around your waist.] These offers are no longer random and are therefore increasingly effective. Leveraging the existing data from your previous activity to predict what will happen in the future is becoming, rightly, more prevalent and valuable to social networks that can sell this promise to businesses and intermediaries.
Do you walk down the same street at dinner time every day? Wish restaurants on that street would compete in real-time for your business?
As social networks add in more location-aware features like Facebook Places and whole new businesses are built on the promise of geo-location including SCVNGR and ShopKick, predictive analytics deliver insights into where groups and individuals will be and when, not to mention what their interests may be. For businesses, there is big money to be spent on location-based advertising in the coming years. As a result, social networks can run their existing location data through predictive models to provide companies with future insights into where to allocate their marketing and advertising budgets for the biggest returns.