Toronto startup Delphia Inc. is developing an investment platform with an extra means of potential profit: paying consumers to share their insights about public companies, markets and their behaviour.
The company will begin taking applications Wednesday from consumers who want to share those insights, with a goal of signing up 100,000 Americans and Canadians in a matter of months. When it reaches that critical mass, Delphia chief executive Andrew Peek says the company will begin to analyze participants’ insights with its artificial-intelligence technology, combining them with other public data sets to forecast where markets are headed to design an investing portfolio. North Americans will eventually be able to access a crowd-forecasted portfolio through a managed investment account with Delphia’s custodian, Apex Clearing Corp.
The more data consumers share – such as by answering survey questions more frequently and giving access to social-media and purchasing histories – the greater a chunk of Delphia’s 2-per-cent management fee they’d be entitled to. These data would, with the aid of machine-learning algorithms, be used to generate forecasts about public companies. And those forecasts would help design a service that Delphia’s team believes would give retail investors a data-driven advantage in a market that is usually available only to high-end investors with deep pockets.
Delphia sees itself as a benevolent alternative to this decade’s dominant data-driven business model, which tends to deliver profit to those who collect data rather than provide it. But the company’s new model is also built on a precarious assumption: that beating the market would be an easy fight. Even the alternative funds that Delphia wants to beat – including hedge funds and quantitative funds designed with advanced mathematical models using expensive collections of public data – struggle to keep up with the markets. In the past decade, the S&P 500 has outpaced multiple hedge-fund indexes, including the Credit Suisse Tremont Hedge Fund Index.
“All of this might work, but it might not – this is still an emerging field of research," said Dan Hallett, a principal at the Toronto-area HighView Financial Group, where he oversees portfolio construction and investment program design. While he suspects it will take a massive number of participants for the service to be accurate – and weed out bad actors and short-sellers – “it will appeal to some people in that they’re offering to share some of the rewards they make from the data.”
But Mr. Peek and Delphia’s investors believe self-reported data will give them an edge over deep-pocketed funds. Janet Bannister of Real Ventures, one of the venture investors who have so far poured US$14-million into Delphia, said that “the richness of the data is something that will be far superior to what other investors or hedge funds have.”
The company believes that with enough people volunteering data – Mr. Peek suggests that a million would be “great” – its machine-learning algorithms could give it a fresh edge in predicting market behaviour. Its co-founder and chief scientist, Clifton van der Linden, earned renown in numerous countries for his previous endeavours in the intertwined worlds of social science and big data. He founded the Vote Compass application earlier this decade, using questionnaires about a person’s political views to determine the party platform with which they best align.
Delphia was spun out of Mr. van der Linden’s politically focused data company, Vox Pop Labs, two years ago, focusing on market analysis as it joined the popular U.S. accelerator Y Combinator.
While quant funds can easily buy data sets with partly inferred insights into social-media or purchasing behaviour, self-declared sentiments about the markets could add an extra layer of clarity to forecasting when building a portfolio, Mr. Peek says. Delphia’s participants will be able to choose to share information from their various social-media and fintech accounts while completing questionnaires at a rate of their choosing.
Aggregating sentiment about the launch of Tesla’s new Cybertruck, for instance, or about events in the world, would help Delphia model its forecasting for specific stocks and industries. The more data a user contributes, the more the “dividend” from Delphia’s management fee. According to the company’s optimistic early projections, a user who only provided data to the company without investing a cent could make as much as US$3,560 over 15 years.
While Delphia will buy publicly available data sets, Mr. Peek says the firm has no intention to ever sell what it collects from users. “We cannot offer the world any value without that data, so it’s critical that we don’t do anything to break that trust,” he said.