Financial services companies have spent billions of dollars and countless hours reinforcing and protecting their data infrastructures in a costly and inefficient manner. As we enter an era that will continue this trend and rely on data more than ever before, we are seeing a new data supply chain emerge to help companies access, store and protect data. At the center of this new ecosystem, key players are developing solutions to make data more operational so that data sets are readily and easily available for digestion.
During the recent “Crux Fall 2020 Summit: Evolution of Key Players of the Data Supply Chain,” 40 leading practitioners and thought experts in the data ecosystem examined the latest developments in data supply chains over a three-day event. Our CEO, Joshua Pantony had the opportunity to participate in this discussion with Emmett Kilduff from Eagle Alpha and Matt Fitzpatrick from McKinsey for a conversation about the future of the data supply chain and expected developments and trends within the next 10 years.
From this discussion, Joshua noted five key takeaways on how Boosted.ai is approaching the evolving data supply chain and positioning Boosted Insights to take advantage of forthcoming developments in data.
Our Five Key Takeaways
Finding Your Role in the Data Supply Chain
The data supply chain has many participants, with Boosted.ai sitting towards the end of the pipeline processing data. This continuous flow of data from various sources feeds our platform, Boosted Insights, which in turn incorporates machine learning to integrate large amounts of data into the investment process and find patterns and problems to make predictions and explain those predictions to users in an easily digested way. Boosted.ai is one member of the chain positioning itself to adapt to the possibilities that emerge with new technologies, such as visualization, interpretation and explainability of data – all things which were far less possible five or six years ago.
Setting Up Supply Chain Factors
In building a supply chain, fundamental and quantitative managers approach the process from two different ends of the spectrum. Fundamental managers have the inherent need to understand why the system functions as it does, and quants tend to only seek to understand how to build a strategy and evaluate associated risks. With the additional processing Boosted.ai provides through machine learning and data, fundamental managers are able to increase their understanding and confidence in buys or to explore buying and selling options that they hadn’t considered before. On the quant side, managers no longer have to operate in a black box. Boosted.ai illustrates their systematic results and includes the factual proof points that led them there.
The True Value of Data and Modelling Risk Factors
Data is viewed as an asset that can predict what will happen in the future; however, this year, traditional risk models broke down and bar risk models didn’t quite capture the impact of COVID-19 in March. Boosted.ai saw this problem emerge once again as the news of a vaccine broke down traditional risk models. In these cases, backward-looking data did not provide additional value and was not able to successfully predict the impact of COVID-19 on the markets. To counteract this, Boosted.ai was able to use varying types of risk factors to ultimately model COVID-19 as a risk factor itself. Boosted.ai sees the value of the data supply chain, but when faced with unprecedented risks, it was humans rather than machines that demonstrated the ability to adapt and successfully emerge from unexpected risks.
Fighting Against Bias in Data
Finding relevant data sets is challenging for a number of reasons, primarily that core problems tend to repeatedly occur. Often, data sets may have a look-ahead bias, survivorship bias, or bias against delisted companies that can influence how a machine interprets and digests the data. Boosted.ai is well-positioned to avoid these biases with our approach to data cleansing and normalization that is focused on both determining and solving for bias.
Benefiting from the Future of Data
Compared to the history of financial markets, the use of data within financial services is relatively new, and a lot of firms are playing catch-up. The need for data differs from company to company, but data offers a clear value based on its ability to make informed predictions. As machine learning evolves, so will its ability to make better predictions. Over the past five years, we have witnessed a massive explosion of data, but even with this explosion, the vast majority of data still isn’t operational. In the next five years, we predict that more and more data sets will come alive and be used systematically. As an industry, we have been laying the groundwork for the last five years, and the next five will be about taking advantage and bringing to life the value that data offers.