Author: Ajith Raam D
Ideas, vision, trust, and that strong “gut feeling” may have helped investors make decisions about how promising a company is in the past, but those days are well behind us. What makes a successful investor, today, is the perfect blend of relying on qualitative information, and quantitative data — which is where we come in.
However, providing them with sheer raw data is not enough — what truly helps an investor in making a sound decision is the ability to understand and manage that data so as to positively influence their investment decisions, be it now, or in the future.
What makes a good investment?
When looking to invest in a company, there are some parameters that are standard. Does the company in question have a unique idea with a strong narrative? How about a business plan? And what about how financially relevant this company is?
Though investors seem to broadly look at the same aspects when assessing the health of a company, their needs are vastly different from one another. Active Learning has a big role to play here in personalizing the needs of each investor – stemming from the notion that one size does not fit all.
Typical data would lead to clues on which company is making breakthroughs in growing sectors, what kind of risks would arise from investing in a certain company, etc. On the other hand, what we offer is a more ‘unstructured’ model. An active learning approach is adopted to study each parameter and customize it based on individual preferences to make insights personalized.
Understanding the investor
Can numbers lead to understanding the personality and preferences of individual investors?
The short answer is, yes.
The process by which we go about this is breaking down the earnings call transcript of companies and passing it through a pipeline of NLP stages that records concepts with their relevance on typical dimensions of interest to an investor. These dimensions are based on the source of news ( CXO, Press, QNA, News..) and position (Quarter, Company, Industry). For each combination of source and position, a relevance score is created for each topic. By requesting an investor to annotate a small set of call earnings transcripts with concepts that he may find relevant, we are able to understand the pattern involved in preferences and hold this information for future highlights for the specific investor.
Picture this scenario — an individual is interested in investing in the auto industry. The pipeline collects call earnings transcripts for a period of four quarters from companies across the peer group (along with other public data such as news, reviews, etc) and knowing his preferences, highlights relevant concepts from the earnings transcript. Additionally, it associates textual and financial sentiment toward the concept allowing the investor to compare across companies, making informed decisions.
What’s notable here is how time-efficient this process is. The investor does not need to go into detail about who they are, what they like and dislike, and what makes them tick. This process sheds light on the personal traits of an investor and how these could affect their affinity towards one company over another.
One investor, many industries
The process we use is further immortalized in its usage across different industries. Let’s say the same investor who wants to dip their feet in the auto industry, is also interested in the tech industry. While the keywords and findings may differ, the preference parameters remain the same for the investor to use in future decision-making as well.
Innovating for tomorrow
Our process helps tap into a much bigger data universe to build a better mosaic around a company. Through this, we are able to comb and filter out the finer details from tons of raw data so the investor can make informed decisions.