Financial Data Science and Machine Learning: the gold mine of financial institutions


The financial industry has a great fear of technology giants: Amazon, Google, Facebook and Apple in the lead. It’s understandable. Big Tech firms have huge masses of users and know them perfectly well, thanks to their ability to analyse data. That’s why we could say that the ultimate frontier of Data Science doesn’t belong to Universities but lies in their hands.
When they will enter the savings and investment market, expect a huge fight. However, traditional financial intermediaries have several weapons at their disposal.


First advantage: data

Basically, every sector of the economy has access to an amount of data that would have been unthinkable a decade ago, and the wealth management industry is no exception. In fact, banks, insurance companies and asset managers have lots of data from which they can obtain high-value information thanks to Financial Data Science, which in our vision includes Machine Learning, Artificial Intelligence, Data visualization and Business Intelligence.


Typically, intermediaries who manage investments have access to the following types of data:

  • financial data, relative to present and past positions: financial transactions of customers and their payment flows – data from which it is possible to obtain information on the dynamics of their investments, as well as on their consumption and saving habits;
  • socio-demographic data, such as age, place of birth and residence, sex, family status and so on, which can be fundamental for framing the client’s investment life-cycle;
  • the answers to the MiFID questionnaire (which, if well designed and filled in correctly, is a great pool of information), that can be crucial for obtaining the client’s financial DNA;
  • customer-intermediary interaction data: such as those related to the use of their online banking private area, the opening of any newsletters, the use of apps, phone conversations, and so on.

Even without additional data (such as those obtained from social networks, or specific smart engagement activities like quizzes, gaming, etc.), it is clear that this represents a considerable amount of information.

First of all, these are very valuable information, mainly because of their specificity: they are related to one’s personal economic and financial sphere. And since we are talking about investments and wealth management, it is obviously much more relevant than the passion for cute little pets or the meme of the moment that can be found on Instagram or Facebook.


…from a GDPR perspective, intermediaries are fully entitled to use the data they own

It is also a data set that can be “increased” – without making big flights of fancy – by cross-checking it with various sources of external data, primarily, data from financial markets, economy and news. Then, if desired, there are various data set alternatives, such as those related to the sentiment analysis, or to the geospatial sphere.

From a GDPR point of view, an even greater advantage is that intermediaries are fully entitled to use these data since they are related to the client’s financial dimension and are used to solve the issues underlying their contractual relationship. However, for Big Tech firms, this element could instead represent a sort of entry barrier.

But how can Artificial Intelligence and Machine Learning help the wealth management industry? There are many applications, and they can all affect the entire value chain. For example, with the aid of data and algorithms, you can solve problems like:

  • identify and understand the financial needs of customers and their real goals;
  • build predictive models of customer behaviour (e.g. whether they are going to buy a given financial product or not);
  • identify customers with high growth potential;
  • segmenting customers, offering them tailored investment solutions and other additional services, with the result of improving their user experience at very low costs;
  • identify which financial needs are met by a given product, and for which purposes it is recommended;
  • support the financial advisors, agents and other relationship managers with targeted customer information and recommendation systems that can tell you which solutions are the best to offer to the customers, based on their specific needs and features;
  • manage compliance in real time;
  • mimic the impact of market events on processes, masses under management, costs and margins, by implementing what is probably the most useful form of risk management for a wealth manager;
  • capture and analyse new data sources.


…financial firms can obtain extra value from a variety of internal and external data sources

Therefore, financial companies can have a significant added value from numerous internal and external data sources, using Data Science tools, which can, in turn, help the strategic and daily operations of all the company’s departments.


Second advantage: data quality

Data is the raw material, no doubt. And their quantity is important, but their quality is far more relevant. In terms of business, data are crucial: regarding Machine Learning issues related to asset management, it is rarely a good idea to indiscriminately throw data into supervised or non-supervised algorithms.

In fact, the financial sector is highly regulated, with professional procedures that own a strong and well-established logic: it would be crazy to not include this “structural” information in the number crunching process. The feature selection is crucial if you do not want to have algorithms that work well in the training phase but are incomprehensible and work poorly in everyday operations.

In our opinion, the understanding of the business behind and around data is far more important than the technical solution, i.e. the development of complex, but end to themselves models, must be taken into account in the development of the very same models. In our experience, large black-box models tend to overfitting and data snooping: these terms mean that the model has not really entered the logic of the problem it wants to solve but has just worked out a very sophisticated interaction with little predictive value.


… the understanding of the business behind and around data is far more important than the technical solution

Furthermore, for many algorithmic applications related to asset management, regulators want to be able to make the look through, i.e. open the model box to understand the logic and the causal links. In these cases, it is not a brilliant idea to carry out an analysis via deep learning, throwing all the possible information into a black box.
That’s why domain knowledge is fundamental if you want to use Data Science for concrete, measurable actions, with a high ROI on investments in technology. This is a huge advantage for financial intermediaries.


Data that generate value

Financial intermediaries already own data of tremendous value, both in strategic and operational terms. But it’s time to exploit these valuable pieces of information with wisdom and practical sense, by employing the tools offered by Financial Data Science, i.e. Machine Learning algorithms and everything that falls under the broadest Artificial Intelligence umbrella.
To sum up, Financial Data Science triggers efficiency and scalability. Which in turn translate into greater productivity. In other words, better margins. These days, it doesn’t sound bad at all.


Virtual B Fintech solutions

Virtual B has been working for years in the financial sector, with a close focus on data and data analysis. Our experience has led to numerous solutions that generate value and solve issues for financial and insurance intermediaries.

SideKYC® is an advanced data analytics software created by Virtual B for banks and insurance companies. SideKYC® can profile customers, identify individual needs and map them with the best product.

If you want to find out more, visit the page and contact us

More on SideKYC®

Download our free white paper “Wealth Management and Financial Data Science: a short guide”

Leave a Reply