On Tuesday, January 15, 2013, I attended the monthly NY Enterprise technology meetup at Cooley. The featured demos were eZ Systems, a globally recognized commercial, open source content management software provider; Pursway, a big data algorithm developer that identifies social networks of customers and key influencers in that network; MemSQL, an in-house memory database for real-time analytics founded by former Facebook engineers; and Visual Revenue, a platform that helps media organizations and editors set the best possible front page using their content and layout.
Visual Revenue decided to “participate in editorial decisions to carry stories on the front page,” Dennis Mortensen, CEO and Founder of Visual Revenue, said. “We came up with a platform that, at its core, uses predictive models that can predict half an hour into the future.” Visual Revenue maps out the value of the front page. “It takes up tof 14 days to train in data and learn the value of the hero spot,” Mortensen said. The platform can translate the value of the front page by analyzing the direct click-thru rate, which can predict the rate of clicks per hour. Data is analyzed in real time. “You need to make sure you align stories with taste of the audience,” Mortensen said. “Make you align with the map set in place.”
Visual Revenue delivers a first-hand understanding of the consuming audience—and their interactions. It has the capabilities to recommend that editors change stories on the front page. It can edit headlines and test them in real time. The test run splits into to experiment groups to analyze which headline is more effective. “The window of opportunity is short, so [Visual Revenue] capitalizes on it in real time with two simultaneous experiments,” Mortensen said. “We have the same quality control as Google,” he said. It is however, very different than traditional analytics. “We aren’t trying to develop yet another analytics platform,” he said. “Editors need input on what to do next, what the most valuable action they can take is.” The platform also gives the user an option to snooze the recommendations from a 30-minute setting, to a 60-minute setting, giving editors the power to choose.
The company is “not here to maximize our cost. We make sure that it learns the data of the publisher’s content, but can be injected with recommendations from editors to change it,” Mortensen said. Visual Revenue charges by volume and it is currently “just focused on content.”
MemSQL, developed by a group of former Facebook engineers was demoed by Nikita Shamgunov. MemSQL is an infrastructure company built by database engineers. “It’s a fairly new company,” Shamgunov said. The company employs 15 people and was founded in 2011. “Our vision is to speed at scale,” he said. “Big data is only useful if accessible,” Shamgunov said. “SQL is a must-have for analytics. Our inspiration for MemSQL came to us while at Facebook. We realized the time value of data. For example, clicks, tick data, machine data, impressions, are all difficult to query real-time and historical data simultaneously. MemSQL lets this to be combined, lets users to go back in time to search.”
MemSQL is able to consume millions of events per second even while querying data. “Memory matters,” Shamgunov said. “It is faster than disk.” According to Shamgunov, data is an inflationary coefficient in the enterprise and that CPU is going multi-core. Storage, especially memory, will be dropping in price as well.
MemSQL’s terminal shows a box for visualization. The demo ran on 32 machines simultaneously. The visualizations showed the engagement of all 32 CPUs when prompted to do something by Shamgunov. “The value of MemSQL is that users know instantly what is going on,” he said. “MemSQL is also compatible with MySQL, noSQL, Ruby…it also has its own database for indexing purposes.” MemSQL can handle as much data—or number crunching—that is pushed through it.
EzSystems, a web CMS Development Company, was founded in Norway in 1999. The platform works in a four-step cycle: Deliver (personalize, recommend, engange), Optimize (learn, adapt, convert), Automate (content creations, marketers, information workers), Create (author, capture, collaborate), and back to Deliver. These processes are optimized by EzSystems. It provides real-time analytics and e-commerce is a key point in the platform. EzSystems tries to make an easy, streamlined web experience. The user—the content manager can use simple tool box to create content or articles. It allows easy uploads (documents and pictures) and the creator can add locations (using Google Maps), tags and pictures to the content. “It’s very straightforward,” Roland Benedetti said. “From here, you need to deliver on mobile, social, the front or the home page.” There are tools for editors to allow them to choose where the content gets published. Editors can also edit the home page. EzSystems provides a scheduling system that can alter the layout and the option to view the scheduled articles in the future.
The editors can also view the analytics side of EzSystems. Visualizations of article views, hits, clicks per hour are displayed on an eye-catching dashboard, which also present which articles are doing well, versus which are doing poor. The platform also updates in real-time and the editors can view where traffic is coming from, whether it’s from search engines or from direct links. It also can recommend content. “The point is to give an editor a very quick and fast way to see what is working and how to react,” Benedetti said.
Pursway, presented by Elery Pfeffer, uses proprietary account big data algorithms to identify existing purchase influencers on your customer base. The platform takes the identified influencers and converts them into a PIVO score. According to Pfeffer, the influencers that can convince others to purchase products are not high spenders, high net providers, highly active bloggers or high spenders. Instead, they hold high influence in small circles of friends. “Purchase influence is topical,” Pfeffer said. “People look in their social circles for experts on a particular topic. It’s about trust and knowing that they want to help me.”
The PIVO score effectively ranks people according to their purchase influence. What Pursway tries to do is sort out and find out how people “act together” in events. “It leaves a digital mark,” Pfeffer said. “As these links accumulate, [Pursway] finds links between networks. Once enough links have been established, it knows that certain users are connected. This is the basis of finding a network.”
The PIVO score reveals the true relative value of customers. “PIVO score generates significant increases in marketing ROI [Return on Investment] be harnessing for the first time, tangible social influence, which companies can now use to increase revenue,” Pfeffer said. Pursway works through annual subscription models. Pfeffer revealed that “retailers aren’t interested in how many friends a consumer has, or what their traffic is to their blog. Retailers are interested in driving revenue, pure and simple.”