By 2020, it’s estimated that 4.4 terabytes of data will be generated each month for every human on earth. From cell phone calls to social media, credit card purchases to satellite images, the human race is creating data at an unprecedented rate.
In every sector of society, from space travel to healthcare, corporations are scrambling to make sense of the vast amounts of data being created, and figuring out new ways to use them for boosting their business. Financial markets are no different.
In asset management, this data goes far beyond the kinds of information traditionally used by investment managers in day-to-day operations, such as market data from exchanges, regulatory filings, and earnings reports.
To get an informational edge, asset managers are now analyzing more unconventional sets of data such as credit card transactions, social media trends, patents, satellite imagery, geolocation information, shipping container receipts, supply chain data, company and product reviews, location data from cell phones, public records and website scraping. These are just a few of the types of data that asset managers are talking about publicly, and it is likely this is just the tip of the iceberg.
| “A recent survey found the most popular source of untraditional data among the buyside are social media posts. They offer a direct sentiment about how a stock or product is viewed by the public, which could potentially hold insight into how fast a stock’s price will rise or fall.” |
With the growth of systematic passive investing, active asset managers have been under immense pressure to demonstrate their worth in the digitally-minded world. Seventy-eight percent of fund managers expect to use new kinds of data going forward.
Unlike traditional data sets which tend to be numerical and structured in an easy to read format—i.e. columns and rows in a spreadsheet—many of the newer data sets are unstructured, qualitative and in many cases extremely large, earning them the moniker “big data.”
Often the data is so big and messy that the analysis itself becomes cumbersome and costly, requiring expensive Artificial Intelligence(AI) technologies like Natural Language Processing(NLP) and Machine Learning(ML) to make sense of it—meaning that smaller investment shops are most likely to be late in catching up.
The data is being used in a whole range of departments, from risk management to the back office, but the leading application is finding new sources of alpha. Advocates of untraditional data say that they contain gold nuggets of information not already included in the market price, which could give an edge in predicting how the price of a security will rise or fall.
However, many asset managers are finding that any alpha contained in non-traditional data is eroded as soon as the data set is common knowledge. By the time everyone on the Street has the data, it loses its value—which is why many asset managers are keeping their sources of untraditional data a secret.
Even with these challenges, the data boom is far from over. AlternativeData.org, a trade body for the industry, forecasts that expenditures on collecting and managing unique data will climb to over $1bn by 2020.