Global Predictions Raises $2 Million to Empower Investor Decisions through Macroeconomic Models

Mitigating investment risk through insights to local and global economies has been sealed away in proprietary software at powerful hedge funds like Bridgewater Associates and BlackRock — until now. Alexander Harmsen and his growing team at Global Predictions are working hard to unlock access to these complex and insightful macro models for mainstream investors. In side-by-side comparisons, Global Predictions is outperforming the IMF World Economic Outlook report’s forecasts by 34%.

The company, which created a dynamic model of the global economy by mapping millions of socio-economic relationships, has announced a $2 million round from Village Global, Morado Ventures, and AME Cloud Ventures. The funding will accelerate the processes of bringing Global Prediction’s tool to investors and of talent-based hiring to enable further growth.

Alex, an On Deck Founder Fellowship alumnus and current On Deck Investor Fellow, spoke to us about the past, present, and future of Global Predictions.

Thank you for taking the time to chat with us! Global Predictions isn’t your first stop on your entrepreneurial journey, can you take us through what has led you here?‍

‍

Alex: My education background is in economics and physics. I’ve always seen the world through models. It shapes the way I make decisions, interact with people, the way I run organizations — essentially trying to take complex information and break it down into specific, much simpler models, to be able to analyze information to make decisions. 

I really loved diving into software and AI, which led me to spend time in Menlo Park. I was the first software engineer at Matternet, a super early stage startup deploying robotic systems all over the world, including aid and medical supply delivery in developing nations. I also had the opportunity to work at NASA’s Jet Propulsion Laboratory on some incredible AI software systems for Mars Helicopter, Ingenuity.

These experiences led me to start Iris Automation in 2015, a leader in autonomous industrial drones and the first autonomous vehicle company on earth to get regulatory approvals and sell products at scale. We grew the team to 55 individuals and raised over $25M in capital from top investors like Y-Combinator and Bessemer Venture Partners. At its core, we were trying to simulate the world and let a tiny microprocessor see the world the way humans do — to figure out how pilots see the world and translate that into a model enabling autonomous drones to react to any potential scenario in real time.

I spent a significant amount of time working with my partner, Jo Varshney, DVM/PhD’s company VeriSIM Life, an incredibly ambitious and successful company building an AI-enabled digital twin of the human/animal body to be used in drug research and development to de-risk clinical decisions at an early stage of the drug cycle. 

In building the architecture and developing these hybrid-AI complex models, we had to ask a lot of questions about how to create the structure, find the right optimization functions, integration points, build explainability into the models, and focus on making the right assumptions about the world we were simulating. 

Iris Automation and VeriSIM Life are obviously in different domains, but these are the same questions we’re now tackling with Global Predictions.

You made the leap from building a digital model of the human body to one of the global economy. Which one was harder to build? 

Alex: That's a good question. Let’s break it down.

I’m not sure I can say which one is more difficult. Both are complex systems with a significant amount of intra and inter-system variability. I do think VeriSIM has shown me how it’s possible to break down extremely difficult biology into smaller problems using first principles.

I've done a bit of quant trading and building algorithms to play in the stock market over the last decade. There is so much data and analysis out there, it feels like the largest problem in the economic space is removing bias, making the right assumptions, and cutting through the noise. It felt like everyone talks about this type of unified, global model [the one Global Predictions is building] and agrees that this would have a massive impact if built properly. 

Yet, most macroeconomic insights available are narrative-driven; in other words, they try to rationalize & simplify what's happening, rather than really trying to understand what's happening. Humans crave narrative to explain what's going on in the world. This is what the media caters to — the idea that people want a story. I think there's an amazing opportunity in trying to remove narrative bias from decision making.

‍

How did you gain conviction that you could build this model and that it would be your next venture?

Alex: After speaking to about 400 people in the space over the last two years, it seemed obvious that this type of widely available global model was an identified need. I really started questioning why no one had built one before. I felt like there exists a roadmap to be able to build it, so what was I missing? That became a fascination for me in 2020. I had a goal to find out what the barriers were.

I didn’t feel like I got satisfying answers as to why. The answers mostly centered around: “it's too difficult,” “you'll never be able to predict all the Black Swan events,” and “it’ll take a billion dollars and all the experts in the world to be able to come together and actually build something like this in software.” 

However, there were a few people I spoke with that said, “Well, Bridgewater has done it. There are a lot of insights you can learn from them, they started 15 years ago with much less advanced technology.” That, I think, is what I'm chasing now — we're not trying to build the perfect oracle, we’re building something better than the currently available status quo — I actually think that’s worth emphasizing.

For a project like this, you just have to start and dive in headfirst. You have to put an initial prototype together — evaluate properly, understand what the baseline is. Then you create an optimization framework, and iteratively work to reduce errors to get to a point where you match and exceed current human judgment. That’s surprisingly within reach.

The world is so complex, and there's so much information out there to synthesize. You really need to automate that to reduce the room for human error in the analysis and decision making process.

It sounds like you’re really thinking about this model as a way to avoid biases that happen when people try to explain instead of understanding what’s happening. Ultimately, the complicated model you’re building will be used by people, how does that become a user-friendly tool?

Alex: The core of Global Predictions is a massive data lake with hundreds of streams of information pouring in every week. Using that, we’ve constructed an economic map, which models out millions of directional, weighted relationships within the economy. On top of that, we have mixed models that are used to run simulations and generate our forecasts.

Around this core, we are building application-specific businesses. The first business is a portfolio management tool, which we’ve been beta testing for about three months. It allows users — medium to large individual investors, family offices, hedge funds, pension funds, wealth managers, registered investment advisors — to take their portfolio, connect it into our portal, and receive a ton of different insights specifically curated for their portfolio.

For example, the tool breaks down whether certain sector exposure will give an expected performance of the portfolio and ties that to underlying correlated drivers for each of the different portfolio items, whether that's ETFs, various public equities, or a number of other securities that we’ve modeled out. Then we'll show a breakdown of what they're connected to on a much larger economic map. Based on the different risks exposed, it will automatically suggest different ways to increase diversification or even increase ROI in a portfolio. For every single one of those items that we're exposing, the user will receive an explanation — a view into the knowledge graph to explore the models in order to truly understand. 

‍

Alongside this portfolio management tool is a general Economy Exploration portal - it allows the user to dive into economic factors, commodities, economic anomalies, public health data, and more. You can think about this portal as a global dashboard for the past, present, and future.

‍

Can you tell us a bit more about your fundraising process with Global Predictions?

Alex: Honestly, I feel like every fundraise I've done is different. I think this time around I wanted to be very intentional and make sure I set up properly before I took in any sort of external investor money. I wanted to be set with the first version of the product, I wanted to build an incredible advisor team. I’m very proud of the people that we pulled around the company. I also wanted to have a solid base of early beta users, and the go-to-market strategy and financial plan sorted out.

Once I had all that, I spoke to many people and investors — not asking for money, but with the intention to understand what they might want to see from the product itself, because there was an overlap between the investors I want to invest in the company and potential customers

Initially, I was telling people that we weren’t actively fundraising yet. Then I got an offer from Erik Torenberg and Anne Dwane at Village Global and I couldn’t resist. They invested in VeriSIM Life, fully believe in the long term vision, understand the value in building out these models, and are cheering for the wide-reaching impact this will have on society. Once I had their backing, I decided to go ahead and get started with the process.

This was really crucial timing for me as it was about three weeks before my wife was due with our first son. I told investors that I was looking to raise approximately $2 million and that I was going to stop on the day my son is born. That had a strange effect on the round — in that it lit a fire. Everyone made very quick “yes or no” decisions. Which is what you want, right? As an entrepreneur, you don't want to be in a very long process as it is distracting for the business.

So much to celebrate — the birth of your son and closing the fundraise! Around the same time you were also starting the On Deck Founders Fellowship. What brought you to On Deck and to the community? How has it been valuable?

Alex: I see a lot of value in curated networks of ambitious, interesting individuals. I went through Y Combinator with Iris Automation — being part of that network was, is, and will be extremely useful. The college network I am a part of has been invaluable. I am also part of Loran, a large scholarship network in Canada, which has been incredibly useful. I saw On Deck as another highly-curated community of motivated, interesting individuals to join.

I like the way it's set up, in that there's almost a real-time pool of feedback that is available. People ask open questions or have very direct queries on something very simple like, “I've never dealt with this tool before, can someone hop on a call and show me how to get the most out of it?” I’ve done that multiple times now and have gotten realtime help from other Fellows, for example, using Mixpanel, Intercom, and some other SaaS products. 

There’s value in getting connected. I've had a number of conversations about people joining us as Head of Business Development, Head of Quantitative Economics, and our other open roles. There was a fundraising concierge program that I took advantage of and got introduced to a number of different investors through the On Deck Angel Network. I even got feedback on the pitch and the deck itself.

I’ve also recently joined the On Deck Investing Fellowship in the hopes that the participants will become 130 users of Global Predictions, because there's a significant number of people managing a portfolio that we might be able to help understand the world. A lot of my initial conversations with folks in ODI end up with questions about how to get access. We have a freemium offering — anyone can go on and sign up, take it for a spin, upload their portfolio and gain insights from the platform.

We’re excited for you to take ODI by storm and for ODI Fellows to get to interact with the product. Let’s end with looking into the future of what this funding round means for the coming year and the vision beyond.

Alex: The current focus has been on harnessing feedback from early beta users and launching the portfolio management tool. We’re ensuring that the tool is as useful as possible, striking the right balance between accuracy and explainability. Naturally, we’re continuously iterating on new versions of the economic map, the models themselves, and the automated ways to extract insights for users. 

The rest of the year will be all about building out the team properly and moving from v1 to v2 on the core tech and product. We’re very set on keeping the bar high using a form of talent-based hiring.

We are going to continue iterating on the core technology, and work hard to make it more accurate and representative of the real world. Around this core we will build application-specific businesses to increase its impact. While portfolio management feels like the most obvious first application, we're also working on building out an API — allowing users to have access points and host themselves or use internally based on their needs and systems. Over the next 5 years, we’ll be addressing use cases in governments, major banks, hedge funds, insurance, and multinational corporations. We think Global Predictions can be useful to anyone that's making decisions in the real world. 

There's so much that you can do if you have a better model of the world. The reduced risk increases efficiency in investments and decision making. Ultimately, I think that's the mission here — to de-risk and reshape how crucial decisions get made around the world.
‍

//
Global Predictions is hiring for numerous positions (Head of Quantitative Economics, Head of Business Development, Senior Product Manager, Data Engineer, Full-stack Software Intern). To learn more, visit here.