“You can have data without information, but you cannot have information without data.” — Daniel Keys Moran
In today's highly competitive financial industries where companies started using big data analytics to help make decisions instead of simply relying on experienced professionals. One of the main reasons is because we can find trends and patterns from the data that could explain some underlying truth that’s been going on for a certain period. The next question goes to how to incorporate the right analytics into a certain industry and eventually make beneficial impacts within a company.
This article will mainly focus on the financial research distribution industry and how data analytics could play a major role in it. Before getting into the possible implementations of data analytics or data science, we want to understand how the industry/business works. The financial research distribution companies serve as the middleman between report publishers who can be an individual analyst or a group of analysts, and the readers of the report. In general, research distribution companies authorize analysts a platform so that they can write out and publish a report, the distribution side will then send out these reports in a form of emails or internal channels to any registered readers all under the same platform. Data such as registered names, email addresses, login info, have been collected when readers opt-in any services they purchased. In this case, data analytics could play an important role in helping management to make decisions or making relevant predictions based on existing data that they collected for years.
Building Strong Linkage With Readers
Since readers are the major source of revenue in most financial research distribution companies, it’s important to understand readership patterns. In other words, we want to know what readers prefer and how to keep them satisfied or even excited in the long-run. One potential implementation of analytics is to analyze past data and create trend analysis to show a clear summary for an individual reader. More specifically, we know from the global industry classification standards(GICS) that there are 11 different sectors, 69 industries, and 158 sub-industries. Here, each company will be assigned to one of these subgroups based on their principal business activities. Similarly, each research report that’s published will also be assigned to one or more sectors. When readers are reading certain reports, we would be able to find out useful insights such as their preferred sectors or topics, how often/long they read a report, and what device they are using to access the report. These insights will be highly valuable when companies are sending out emails or reaching out to acquire new customers because they already know the user behaviors and understand well of user demographics. One further implementation of data analytics is to build recommendation systems for each reader so that it could drive the reader’s interest in reading more reports. However, risks that are associated with these methods are the usefulness of the original data because the analysis will require a clean set of data with relevant variables to be able to proceed.
Predicting Web Capacity
There are sometimes where we can’t go on to certain websites because the website crashes. Website crashes and stops serving data usually means that the website has reached a limit of serving too many concurrent users. According to a research paper called ‘Capacity Planning for Web Services’, management usually faces two major challenges regarding web capacity, one is to meet customer expectations in terms of service quality and the other one is to keep IT cost under control. These challenges are also related to the financial research distribution industry because companies rely on a single platform to allow for report publishing as well as enable readers to read reports. The situation where companies don’t want to face is when emergencies occur which may lead to increased tractions on the platform but the capacity reaches its limit so it stops serving incoming customers.
What if we could predict the expected capacity usage ahead of time? Data Science would be able to handle this kind of problem by using machine learning models to analyze the readership trends based on past data and live market data. Fit a model that can help to predict server usage will help companies to adjust the limit on time to serve enough readers while keeping the server cost under control. One risk that may arise is that we still need human intervention while making predictions and should not solely rely on algorithms because there might be computational biases in the data itself. Having experienced professionals to evaluate the prediction results would be more useful in this case.
Data can be stored in many different ways to improve efficiency. According to a research paper named ‘Big Data Storage and Challenges’, there are three main characteristics of big data which include velocity, volume, and variety. Data that’s been stored by research distribution companies can easily fit into these three categories. There’s a large variety of data ranging from numerical to categorical variables that’s available in the databases and live data being generated while a reader is opening up a report or scroll through pages. Here, we can tell what foreseeable challenges a company may face which is how to efficiently store all the data in a cleaned manner while people can quickly locate and extract any information from a huge database. Currently, some companies in the research distribution field are still using hard drives to store all their data which causes the messiness of the entire database. Data migration would be a better choice for companies that store large amounts of structured or unstructured data. Data analytics and structural designs will be necessary to transfer unstructured data into a more structured manner in the cloud. Another benefit that data migration brings is that cloud computing could handle much more complicated computations compared to regular computers which can also bring work efficiencies. One risk with this is that once data has been processed and transferred, it can be difficult to change it, so it’s important to design the data structure before migrating the database.
To summarize, there are many potential opportunities for financial research distribution companies to utilize data analytics or data science to help improve work efficiencies or make the production cost under control. Companies should find their best way to practice data analytics based on their data structures and business goals for better results.
Padgavankar, M. H., & Gupta, S. R. (2014). Big data storage and challenges. International Journal of Computer Science and Information Technologies, 5(2), 2218–2223.
Almeida, V. A. (2002, September). Capacity planning for web services techniques and methodology. In IFIP International Symposium on Computer Performance Modeling, Measurement and Evaluation (pp. 142–157). Springer, Berlin, Heidelberg.