Hedge funds and investment banks are continually challenged to reduce risk and increase alpha while reducing costs. As a result, it is believed that over 82% of hedge funds use some type of alternative data. Alternative data refers to data that is not traditional data such as financial statements, management presentations or SEC documents. The goal is to use alternative data to create an alpha advantage through the development of better forecasting or predictive analytics before their competition.
To meet the need for alternative data, numerous data resources and platforms for finding and accessing data have been launched. Most platforms act like grocery stores with as many datasets as they can find. Many datasets go unchecked and often contain data that is unstructured, unusable, and even non-compliant with privacy regulations. Users must spend considerable time verifying and analyzing the data and must assume that it complies with various consumer privacy regulations.
An alternative data platform that paves the way for cleaning up this confusion is the Bloomberg Enterprise Access Point through its data catalog. Ensuring that all data is verified, compliant and organized for user access is fundamental.
I asked Phil Rist, Senior Vice President, Strategic Initiatives at Prosper Business Development, to discuss the evolution of the alternative data market and why he thinks Bloomberg Data Market Place is one of the leaders.
Gary Drenik: What are some of the challenges that alternative data buyers face?
Phil Rist: First, the rapid expansion of the alternative data market has resulted in the availability of data resources and providers. There are currently over 445 data providers and the number is growing. By comparison, in 1990 there were only 20 salespeople. Second, you need to recognize that not all data is created equal. Once you realize this, you begin to understand the immense challenge faced by alternative data buyers / users who must separate the wheat from the chaff. Determining which data can provide a real lasting alpha advantage is time consuming and tedious. For example, many have used transactional data such as credit cards and others have experimented with web scratching or social media. While each of these data sets can sometimes have some degree of value, none are truly representative of a market as a whole. Additionally, they represent a historical view that is one-dimensional and requires several unverifiable assumptions due to the nature of their origin, click or swipe. They also require a significant investment of time to verify and analyze. Finally, Covid-19 has disrupted the value of time series for many of these datasets and their algorithms. Rapid changes in consumer behavior have created a new consumer market not represented in old transactional data sets. The basic all data guide for all scientists engaged in MLops is still the same â¦â¦ GIGO. Having more low-quality data does not lead to better results. Building predictive analytics that work over time is more important than ever. Several years ago, Gartner said big data was in the midst of disillusionment and that this opens the door to smaller, more accurate, and more representative consumer data sets.
Drennik: What types of alternative data do you see as the market begins to mature and look for more meaningful data to create predictive analytics.
Rist: Several years ago, an article in Wired magazine questioned whether Big Data had passed Gartner’s “hype cycle” and entered the “”hollow of disillusionâ. The next phase is where second generation products start to evolve and innovation becomes more important. Likewise, the next level of alternative data use will be the use of more scientifically collected consumer research data. Investors need data representative of the entire consumer market, do not require any unverifiable assumptions, come directly from consumer responses and follow-ups. consumer intentions. Good quality, scientifically collected, syndicated consumer studies are rare. I know of some sources that take the time and risk of sourcing alternative data and provide accurate analysis available ‘off the shelf’. This type of innovation will help increase the adoption of alternative data.
Consumer-based data has proven to be accurate in predicting turning points in the economy long before the data was released by the government. In addition, data on consumer intentions is useful for forecasting income. We have many use cases where consumption data forecasting events such as the 2008 economic slowdown 8 months ahead and better revenue forecasts for specific brands.
Since consumer spending accounts for 70% of economic activity, using granular data collected scientifically directly from these agents makes sense. We have partnered with the NRF since 2003 as the source for all of their data on holiday spending and consumer trends. So we know that data works as an accurate predictor of spending.
Scientifically measure the behaviors, mood / feeling, motivations and intentions of 7,500 consumers each month and publish them within a week of collection analyzes can be developed 28 to 48 days before government data. The data is complementary to government survey data such as consumer spending surveys, BEA and Federal Reserve data. The integration of this historical government data with consumer intention data sets holds great promise for new perspectives.
Drenik: You recently announced a new relationship with Bloomberg Enterprise Access Point. How do you think this will benefit investors?
Rist: Prosper has evaluated several alternative data platforms over the past two years and Bloomberg seemed more in line with our view that smaller, more consistent, relevant and structured data sets are better than more data of unknown quality. Bloomberg’s due diligence is extensive and the platform is limited to single data. With investment firms focusing more on increasing alpha while managing costs, quality consumer data offers a solution for better predictive analytics solutions, and Bloomberg is enabling it.
Drenik: Phil thank you for your compelling ideas on the growing and evolving alternative data market.