Ninja Metrics has officially launched its Katana Analytics Engine for tracking the monetary value of social influence. Measuring an influencer’s impact down to the dollar is a bold claim that would have an enormous impact on marketers if true, so we chatted with CEO Dmitri Williams to find out more about the technology.

At launch, Katana is designed to work for game developers and marketers, but it will spread to other industries over time.

The idea behind the engine is that it can learn to predict which users tend to have an influence on sales within their networks, both online and offline. Katana then assigns a Social Value in terms of dollars to users. Once developers have identified high-value influencers, they can then customize their ads and send in-game offers to them.

The process sounds reminiscent of social influence startups like Klout and Kred, but Williams asserts that’s not the case.

“We’re different,” he said. “We don’t go directly to the consumer. This isn’t a public score. You won’t have your Ninja score be known. This is a b2b operation. The reason why companies tend to be comfortable with our technology is really the difference between abstract and concrete.”

Ninja Metrics, which just closed a $2.8 million funding round, is banking on the fact that marketers would much prefer to have real dollar projections instead of arbitrary numerical scores.

“Twitter-based systems are good at finding people that when they talk other people listen. Data science says there’s a huge gap between talk and action and we’re into action – dollar spent, shows watched. Whether they talk about it is largely irrelevant.” Williams continued.

Ninja Metrics doesn’t care whether your neighbor talked to you about his new car, just whether his purchase caused you to buy one of your own.

“We don’t care whether you saw the car over the hedge, whether you tweeted about it. We know you’re connected to this person, they did something and then you did it,” Williams noted.

Measuring influence, however, requires an understanding of the data that moves from correlation to causation. Katana uses a proprietary machine-learning algorithm to arrive at confident results. The engine combines developers’ social graphs and transaction data in order to work its kung fu magic.

Williams claims Katana’s predictions are upwards of 85-90 percent accurate on dollar values. Considering that Ninja Metrics estimates that 10 cents to 40 cents of every dollar spent on gaming is driven by social influence, Katana has a massive market available to it.

Image credit: Shutterstock / lijansempoi