A portfolio manager at a $43 billion investment firm explains firms' flawed approach to using alternative data, and how they can fix it

Mike Chen, a portfolio manager at PanAgora Asset Management, said firms are struggling to find value in alternative data because of their approach.Instead of trying out a variety of data feeds, Chen said firms should first develop a specific investment question they'd like to solve.Wall Street investors have recently voiced frustration over generating substantial returns using alternative data in their trading strategies. Like trying to drink water straight from a fire hydrant, investors are having difficulty sifting through the massive amounts of alternative data providers to find insight that can quench their thirst for big returns.But Mike Chen thinks firms need to completely change up their approach to truly benefit from the potential power of the unique data sets. Chen, who is a portfolio manager at $43 billion PanAgora Asset Management, told Business Insider too often firms are willing to test out different data sets without any sense of direction. Chen said the quantitative investment firm always as a specific investment idea in mind before reaching out to any alternative data provider. He described it as a pull model rather than a push model. Everything comes up with fundamental sensibility. What do we want to address？ How do we solve it？“ Chen said. ”Then you work backwards. What is the technique that you should use？ What‘s the data set that you should use？ We ask the investment question first rather than ask it last. “See more: Wall Street's appetite for alternative data has turned into a gold rush, but investors say there's more dirt than nuggetsChen said by taking a focused approach beforehand, he's much more likely to gain valuable insights from the data he gathers. Speaking at a conference in New York on Tuesday, investors voiced their frustration over the limited amount of alternative data they're able to implement into their trading strategies. Chen said he's not surprised. And while PanAgora's process of starting with a specific investment question isn't foolproof, it gives the firm a better chance of finding useful data.”Otherwise you are just throwing darts at the dartboard with a blindfold on,“ Chen said. ”Even asking the right questions, you still won’t be right 100%, but at least your probability will be much higher. Just trying to aim a little bit before they throw.“Read more: Wall Street is obsessed with alternative data, but Nasdaq's CEO is sounding a warning bellOnce an idea is formed, Chen said the firm typically looks to try and collect its own data to aid the trading strategy. They aren't opposed to buying standardized data sets, he added, but they won't use any investing signals created by the data provider. Once a ready-to-use investing signal has been created by the data provider, Chen said, there is likely too much money chasing the strategy for it to be profitable.The process of creating proprietary signals from the raw data isn't easy, he added, but it's all the more reason to do it. ”If it is painful for you, I bet you it is painful for other people too,“ Chen said. ”Once you do it becomes a competitive advantage.Sign up here for our weekly newsletter Wall Street Insider, a behind-the-scenes look at the stories dominating banking, business, and big deals.