Using Machine Learning for Consumer/Retail Private Investing
How quantitative investment strategies can transform the deal for both sides
Investing in private markets poses many challenges. Unlike public markets where it is much easier to pull large data sets, private market investing is tricky, qualitative and relationship driven. Traditionally, investors start with a phone call, and then move to build relationships in any format possible; going to tradeshows, cocktail parties as examples. Certain industries are much less advantaged towards this format of investing. Unlike tech where a lot of innovation happens in hubs, consumer/retail investing is so disperse as far as where people are innovating. The sheer number of consumer companies that are geographically disperse poses a challenge for investors to identify worthwhile investments. For the entrepreneur, wooing investors across the country and standing out in a crowded marketplace poses challenges for accessing capital and resources. The challenges from both sides of the transaction presents an opportunity to bring data to the conversation to transform the deal, so one or both sides does not leave with the taste in their mouth that they have been undervalued, mis-valued or completely overpaid.
What is CircleUp
Circleup is a data driven investment fund that is using machine learning to track more than one million consumer companies and their data to determine investment decisions. CircleUp is enabled by their machine learning platform, Helio, which collects over a billion data points on CPG /retail companies to evaluate them based on a set of dimensions used to predict successful investments [1]. Data on CPG/retail companies is plentiful and dynamic, with info on where products are sold, prices, skus, distribution channels, what end users think about the product, etc., available in both public and proprietary formats [2]. Helio draws on these data points to develop insights and descriptive traits of companies, allowing investors to create a mosaic of a company. In the short term, Helio allows investors (both their own and others) to follow a more data driven in approach to developing insights on a company rather than a more gut-feel approach.
In the medium term, the goal of Helio is to not only mimic the investor’s thought process, but to add more data and intelligence to the deal than a human could do manually. They aim to do this through both predictive and prescriptive analytics [3]. With algorithms that can process billions of data points, analyze the interactions between those and test copious outcomes, CircleUp can create predictive insights that are most relevant to the company and aligned with the goals of both the entrepreneur and investor. It allows entrepreneurs to stay ahead of trends and move into new markets with more confidence, and investors to provide capital based on future potential with more robust supporting analysis. Prescriptive analytics goes beyond predicting future outcomes, but also suggests actions that benefit from the predictions and can model the implications of those.
What to focus on next
To be able to achieve this predictive and prescriptive goal, I believe CircleUp should focus on shortening the feedback loop and double-down on recruiting and retainment. CircleUp needs to demonstrate success to convince limited partners to believe in the vision. I see challenges for raising capital for funds focused on quantitative strategy in private markets because of the long lifecycle and feedback loop. It may take multiple year(s) out to understand whether certain investments/predictive insights were successful. Looking for ways to shorten the feedback loop or find alternate proxy’s for success will be vital to build a critical mass of investors willing to put their capital into this strategy. Finally, because this is an industry with a lot of money floating around, attracting and retaining talent in both business and tech (engineering and data science) will be key to stay ahead of the curve. If CircleUp is able to demonstrate how quantitative investing can be successful, other firms/funds will start building their own algorithms (many have already), and the fight for talent will become the deciding factor for which machine learning platform outperforms.
How applicable is this
It was quite intentional that CircleUp is focused on first tackling the CPG/retail space with quantitative investing. Do you believe that there will be a shift from mostly manual to fully algorithmic private investing and which other industries are most attractive for this shift? CircleUp has the goal of replicating the decision making process and bring a more holistic approach to investing that goes beyond the human brain. What aspects of private investing do you believe will be most difficult for machine learning and algorithms to replicate?
Word Count: 744
Endnotes:
1 Ryan Caldbeck, “The CircleUp Vision,” Medium.com, March 22, 2018, https://medium.com/@ryancaldbeck/the-circleup-vision-50e44eb900ad, accessed November 2018.
2 Company Fact Sheet, https://circleup.com/helio/, accessed November 2018.
3 Ryan Caldbeck, “Announcing The Launch of Helio,” Medium.com, February 27, 2018, https://medium.com/@ryancaldbeck/announcing-the-launch-of-helio-b06458a27af, accessed November 2018.
This article brought to my attention data capabilities regarding investing that I didn’t know existed, so thank you. The biggest challenge that I see with shifting from manual to algorithmic investing is that investors want to be ahead of the curve, not just getting the same information that all others are getting. Once there is enough data about a company for Helios to form an opinion, will it already be too late? To answer your second question, the networking aspect of investing cannot be replicated, and serves the purpose of assuring an investor of their belief in the leadership team. Because the success of small companies is so dependent on their leadership, I think CircleUp would need to be a tool for investors to use, rather than a replacement for the current process.
Machine learning has obviously disrupted public investing to a large extent. However, CircleUp is one of the few companies to bring this disruption to the private investing market. This is intriguing, primarily because private investing relies heavily on operational improvement post-investment. Given this dynamic, I am curious to see how CircleUp’s investment recommendations co-exist with the relationships and decision-making in the board room. Perhaps as a growth opportunity, CircleUp should see whether it can carve out a role in the portfolio operations space using machine learning in the budgeting and forecasting decision-making process.