Marketers relying on a traditional method for calculating return on investment (ROI) of advertising sources and customers bid blindly on media. The social predictive analytics company Ninja Metrics, the brainchild of a University of Southern California professor, believes it can change that by calculating ROI for customers based on a person's social influence.
Ninja Metrics has been testing the technology with companies focused on in-game advertising. The system calculates what Ninja Metrics CEO Dmitri Williams calls true ad value and identifies the players who bring in traffic, as well as the ones that generate the most and least money. It's done by determining the social value of the player and integrating it into the equation.
Most marketers fly blind. "Marketers spend a lot of money on acquiring users through programmatic ad buying, and then they need to determine whether the cost of that person is higher or lower than their actual lifetime value," Williams said. "In other words, what's the cost of acquisition vs. the lifetime value?"
Ninja Metrics built social network graphs to identify patterns. Take the act of buying movie tickets. The platform would identify one person buying movie tickets and whether that action prompted friends to buy movie tickets. A causal algorithm in the platform can identify this trend with 85% to 90% accuracy. The attribution platform, which identifies leaders and followers, can determine where the money originates and where it goes.
For ads that appear in video games, the Katana Social Analytics Engine in the Ninja Metrics platform ties in-game behavior and player influence back to the advertising referral source, allowing companies to track their ROI for each advertising source. Marketers can see what advertising sources are bringing the most dollars and the most valuable users into their games. The program calculates the exact dollar amount from each player.
"We've done analysis on many of our clients and for about 200 million people over many months and found some of the ad sources are much better than they appear, and others much worse than they appear," Williams said. "We can calculate on the fly the campaigns and sources bringing in more or less value than they appear to be. We're finding the average error at about 10% to 40%."
The error signals that marketers are investing the wrong amount of money to acquire traffic. Some are under-priced and others are overpriced. Marketers might think an ad brings in $100, but it really brings in $50. "It's a huge error," Williams said.
The technology, which supports in-game advertising today, also provides marketers the ability to eliminate bot traffic and duplication from reports.