Quantifying Unstructured Data — Anju Marempudi, WG’05
20 September, 2018
category: Entrepreneur
Anju Marempudi, WG’05
Founder and CEO, EventVestor
Anju Marempudi, WG’05, (India) had an epiphany that led to the creation of his company, EventVestor. It was the first day of Professor Michael Useem’s Foundations of Leadership class. Mike Useem put up a slide showing that 3M’s stock price went up over 15%, adding over $4 billion to the market cap on the appointment of a new CEO. Meanwhile, another company lost its CEO, and the stock price went down by over 10%. (I remember sitting in that same class thinking, “Yep, leadership sure is important.”) Anju thought, “Why can’t we structure that data?”
Anju’s 11-year-old company is performing well, but he’s still striving. I interviewed him by phone on a Sunday night about 9 p.m. When he told me that he needed to take a break to speak to his team, I realized he was still at work.
What do you do?
EventVestor structures all information and developments in the news, that could potentially impact stock prices.
Traditionally, quantitative models are based on stock prices and quarterly financial data, because that’s the only information available in a highly structured format. Everything else, the qualitative events, goes into fundamental analysis and was anecdotal.
Can you define “structured data”?
Most of the news about a company is unstructured. A recall occurs, a merger or a new CEO — and investors consume this news one piece at a time.
We work to give investors the same amount of data structure for an event as we would to financial data. Let’s say the company chooses a new CEO. What is the company name, the date of the change, the name of the CEO? Is it an internal promotion, or from outside? Is the CEO a male or a female? An investor is able to compare the company they are looking at, to our structured data on over 6,000 CEO changes, with one-third coming from outside and two-thirds promoted internally. We can analyze how the stock price responded under these and other scenarios in a flash, especially compared with going back and reading multiple news articles.
So, the events in EventVestor are market-driving events.
Yes, and market-driving events happen every day, every hour. We track not only large events like M&A or arbitrage, but also over 100 event categories — for example, a company announcing earnings guidance, shareholder activism or product recalls. Every time we read the news about a company recall, how do we think about the company’s stock price? If I, as an investor, can get the previous 50 recalls in a similar industry, and find out the stock price 10 days later, and one month later, then I can respond.
We tried looking at many “events.” One finding, which put us on the cover of Barron’s in 2009, was when we performed an analysis on Jim Cramer’s Mad Money show. What happens to the prices of hundreds of stocks when Jim Cramer said, “Buy”? Our analysis showed prices rising before he announced a stock, and a few days later, the stock price started to go down. We stated that you could actually gain by shorting his recommended stocks.
How did EventVestor begin?
On graduating from the Executive MBA program in 2005, I continued my work at Merrill Lynch and later at ZS Associates. I had worked at many companies working with data all my career and reflected that there was no structured data around these corporate events. I decided, “Let me start one.” In 2007, I moved to India, because to collect and organize all this data, we needed to use a lot of human intelligence. I started with 500 companies and 30 events, and today, we track 5,000 stocks and over 100 events types.
Can you give an example of tracking an event — say, a merger?
M&A is an important event category. Is it a friendly deal or hostile? Is it acquiring a private or public company? International? Domestic? An all-cash deal versus cash plus stock? We have no opinion about the data. We just provide the best fuel for the quants. Data is the new oil.
How did you respond to the financial crisis?
In March 2009, at the NYSE, listed companies were cutting dividends and suspending buybacks, and NYSE wanted to know how the market reacts to these announcements. The NYSE market desk was performing manual and anecdotal analysis, and it heard about EventVestor. They asked if we could do analyses of how stocks react to company announcements. It saw that it was a much better way to analyze.
Then we began providing insights and analytics for the 2,000 companies on the NYSE. Today, most of our customers are large hedge funds that have their own algorithms. They come to us to give them data with integrity that they can rely on. We also see a lot of interest from the ‘quantamental’ hedge funds and asset managers.
A quantamental hedge fund uses quantitative models to bring efficiency to their processes, by screening out companies according to certain criteria.
What are you focusing on today?
Today, we are expanding our event coverage by focusing on artificial intelligence (AI) and natural-language processing on a wide array of unstructured data such as SEC filings, regulatory disclosures, press releases and website information. We have some of the largest banks and hedge funds as our customers. We are on the Quantopian platform and Quandl, a marketplace for alternate data.
Can you talk about your perspective as an immigrant?
I have always wanted to do something to make a meaningful difference, while being aware of the risks and challenges involved. The internet boom of the late 1990s, along with the startup culture and the success of many immigrants in the U.S., provided a perfect opportunity to start my own business. My India roots, education in India and a good network of friends in India were big advantages for me when I wanted to set up my financial data research operations in India. That has really broadened my horizon and helped me with a very holistic approach to social, cultural and economic perspectives.
What was the impact of Wharton on your career?
Before Wharton, I had started IntelliBusiness, focusing on data and business intelligence. The dot-com crash in 2000 prompted me to reflect on my strengths and market opportunities and so I decided to acquire a business degree at Wharton before I ventured again on my own. Wharton’s financial focus and my data technology background provided the ideal combination to start a financial data and information business in 2007.
I am grateful that, when I returned from India in August 2008, I reached out to 30 or 40 Wharton alumni in the financial services sector, and 80% responded within 30 days. Then, in mid-September, as I was driving to Boston to meet with a customer, I heard on the radio that Lehman collapsed, and from the next week on, no one was calling us for business. My Wharton experience and the lessons learned from many entrepreneurs helped EventVestor to navigate the financial crisis and evolve into a leading and innovative alternative data provider to investment professionals.
And recently, you’ve been giving back.
Over the past couple of years, I have been volunteering as a special advisor to Andhra Pradesh, my home state in India, helping the state develop a world-class digital ecosystem around leading technologies, including AI, machine learning, cybersecurity and blockchain.
I have also benefited from attending many WCNY affinity group meetings. So, wanting to give back, I started the Wharton Fintech Network affinity group for the benefit of WCNY members. Over 60 Wharton alumni attended the kickoff event this May, 2018. It showed the enthusiasm of Wharton alumni – all three speakers at the fintech kickoff event were Wharton grads: Lance Braunstein, WG’05, Chief Information Officer, E*Trade; William Libby, W’03, Founding Partner, Upper90; Radhika Shroff, C’95, W’95, G’00, of ACCION.
I am planning to host one fintech event every quarter. I am very thankful to the Wharton community for all it has given me and for the opportunity to give back.
– by Kent Trabing