How data-driven strategies can improve impact investing outcomes
DATA SCIENCE is making inroads into the world of impact investing, helping program designers and beneficiaries achieve closer alignment between their goals and strategies. While some are building on models from the business world to correlate different pieces of ecosystems to understand how impact flows, others are attempting to marshal next-generation digital technologies such as blockchain to improve outcomes in areas such as disaster response and land titling.
The Rockefeller Foundation has been designing ways to harness data effectively in order to improve the effectiveness of impact investing. “Data really helps you understand the nature of the problem, and thinking about data ahead of time helps you structure your experiments and your interventions,” said Zia Khan, vice-president of initiatives and strategy at the foundation. “Measuring data helps you prove what works, what doesn’t work, and then you can monitor and scale things up.”
The data movement is infectious. Mr. Khan said he sees “increasing appetite to learn more from some of the countries that have done some breakthrough work,” such as Estonia and India. In particular, he cited India’s biometrics-based identification system called Aadhaar, which has enabled millions of previously undocumented residents to open bank accounts and secure government benefits, education and jobs, among other gains.
THE POWER OF CORRELATION Locus Analytics, an economic research and data analytics firm, has developed a new model to locate, map and study the components of an economy, which allows researchers, policy makers, investors, entrepreneurs and anyone interested in economic data to sort, splice and analyze the information in new and innovative ways, according to founder and CEO Rory Riggs.
The company created the Locus Model, which it said enables “a new method of interactive data analytics for business and economic data.” The model is apt for impact investing because it brings more precision in directing resources and maximizes desired outcomes. “If I could give you transparency to all these marketplaces, you now, from a financial point of view, can think through, ‘Where is the most impact for your dollars?’” Mr. Riggs said. “It is a way to benchmark and grade your investments and make them much more targeted.”
Mr. Riggs explained how Locus harnesses the power of interrelation in data. Much like how geographic mapping uses a so-called “coordinate system,” the company uses location-based services to create “a robust classification system to understand better who you are, who’s around you, and then how to find other people who are like you,” he said. “We’ve taken a production system from inputs to outputs and allows you to figure out where you are in that process.”
For example, the model creates correlations between various functions, such as design, production, sales and transportation in a business setting, to create an “activity wheel.” That framework “captures every activity in the economy, and these activities are the same in Kigali as they are in New York,” he added.
Messrs. Riggs and Khan shared their insights on innovating with data in an episode titled “Driving Data for Impact” as part of a Knowledge@Wharton podcast series From Back Street to Wall Street, where entrepreneurs, investors and policy makers from around the world focus on maximizing impact investing outcomes.
WIDENING INTEREST
Mr. Riggs said one user of such data could be an entity like the Farmers Business Network, which crowdsources farmers, connects them with each other and shares realtime information on crop patterns, new agricultural products and others. “This is going to happen industry by industry” by providing participants greater visibility into their individual roles and their impact across their value chains, he said. “I’m hoping that every impact worker will have access to understanding what their business is and how their business is performed around the world.”
The Locus Model helps organizations locate their place in a value chain and then create their own networks. “It’s a neat way of saying, ‘Once I put my own code on myself, then I can find other people that do that, and we can start building neat networks that allow people to communicate,’” Mr. Riggs said. That model could be extended to country-level planning, too. In what he called a “clinical model” for economic development, policy makers could understand how their economy has developed and compare that with other countries.
“Our codes facilitate your creating a clinical model, because it allows you to build a standardized representation of your community, and be able to compare it to a standardized representation of Detroit or New York or Stockton,” Mr. Riggs said. As it happens, Stockton is a transportation hub for farm areas around California. “If you’re building a merchant hub in an impact country, wouldn’t you like to know about the businesses that have formed and developed in an American distribution center? And start building your models around that?”
Locus has also identified stock indices as users of its tool, where it could correlate companies by their businesses. “The S&P, MSCI, Wilshire and Russell are now running indexes using my technology to group companies that correlate because of function,” Mr. Riggs said, adding that investors following those indices could design and diversify their portfolios based on this data.
Insurance companies also are studying the model to understand correlations between, say, disability problems and specific jobs, Mr. Riggs added. “In the financial world, we’re getting good traction by people realizing they now have a new tool to identify and then manage risk because they can see it more clearly.” He is also seeing interest in the model from companies that want to understand their peers better or map their supply chains.
WORKING WITH GOVERNMENTS Mr. Riggs has been taking his model to governments. Locus recently reached an agreement with the government of Rwanda to start developing survey tools for that country that will allow it to start mapping its community in “a functional mapping context,” he said. “What we’ve told Rwanda, and we’d tell others, is that once I have your system mapped in our system, I will handle the [networks across the] developing world,” said Mr. Riggs. “The promise will be in the impact world. If you let me come in and map your community, I will allow you, then, to compare your community to all these developing world countries.” Locus has already mapped China, England, France, the US and Mexico.
In order to ensure that the data in the networks stays current, Mr. Riggs plans to crowdsource that effort by creating online homepages for users to update their information. “Just like people will update their Facebook page or a Wikipedia page, I’m hoping that people will take ownership of the page they care about so that that page stays current.” His firm is also building machine-learning tools that can read websites and update business summaries. Locus already has information on more than 50 million companies in its database, which together account for about 150 million jobs.
Mr. Riggs described the Locus Model as “the ultimate free trade tool.” It allows users to get “perfect information at any given point in time, so the speed of change in business will become much faster,” he explained. Impact investors would be able to map information on new business start-ups, and businesses could use it to manage supply chains, plus find avenues to sell their products and services, he added.
Research work in correlation led Mr. Riggs to create the Locus Model. A math major in college, he had spent most of his earlier career working with scientists building biotechnology platforms to conduct fundamental research. That work involved using data to understand a disease by matching its attributes with those of other diseases. He wondered whether he could build similar platforms for business with a network of all economic activity based on fundamental attributes of business functions. With the Locus Model, he showed that it was indeed possible.