10 Feb 2021Gery Zollinger

How to make data science work for wealth management

Data-driven insights are a real supercharger for a wealth advisory business. And the main ingredient, the data, is already inhouse and waiting to be uncovered. However, most banks and wealth managers are confronted with the complexity and incalculable difficulties of the task early on in their data science journey. With this blog, we help you bring your high-flying expectations into alignment with the sometimes-harsh reality of data science implementation.

Break down your goals into tangible data science use cases

The importance of data analytics and the potential benefits for financial institutions are undisputed facts. It is, however, important to set clear targets at the beginning of a data science undertaking and to capture them in practical use cases. As a report by BCG [1] revealed, advanced analytics bring the potential to substantially increase assets under management, client acquisition and retention. However, business growth is not the only benefit data science can deliver to an organization. Personalization of services, efficiency gains and reduction in risk should not be neglected.

Figure 1: Advanced analytics benefits for wealth managers (BCG 2019)
Figure 1: Advanced analytics benefits for wealth managers (BCG 2019)

In one of our previous blogs, we highlighted the potential of relationship networks and translated this data science application into use cases to help relationship managers identify new prospects and to support compliance officers fight money laundering and financial crime. These are two valuable examples of how the end product of a data science effort can benefit functions along the value chain of a wealth manager.

Overcome your data science challenges

Once your ambitions are translated into use cases, you are ready to start the data science journey. Many wealth managers envisage this as a straightforward, simple undertaking. But just hiring a few data scientists and installing a data analytics software might not be enough to harvest all benefits for your organization.

Figure 2: Expectations on a data science project
Figure 2: Expectations on a data science project

In fact, wealth managers rarely manage to take advantage of the full value of their data, as applying sophisticated analytics methods and algorithms takes them longer than planned. In our view, the following factors often challenge and delay the introduction of data science in traditional financial organizations.

Achieving organizational commitment

Make it a top priority of your leadership and the full organization to transform your business model into a data-driven one. Lone fighters don’t achieve results in this discipline. Develop use cases in an agile organizational framework, involving dedicated resources from front office, back office and compliance.

Mastering technological complexity

The tools and hardware needed to make use of data science are complex and require significant investment. The costs triggered by the installation and maintenance of these complex solutions are often underestimated by financial organizations.

Finding data scientists

Data scientists remain one of the most in-demand job profiles and many recruiters underestimate the challenge of finding people with the right skillset and experience.

Controlling infrastructure costs

The data centers of many financial institutions have severe space restrictions and so their operators opt for more sophisticated and expensive space-saving servers. The requirements of big data software are likely to increase demands on the infrastructure and drive costs exorbitantly when not actively managed.

Cleaning up messy data

The effort needed to understand the oddities and nuances of incomplete, messy data in order to clean it up is said to be the biggest pain point in the life of data scientists. Banks are particularly prone to bad data quality, especially when client profiling information was tracked in free text format over decades. The challenge to make this data ready for usage in advanced analytics is often only to be overcome with the help of machine learning techniques.

Figure 3: The reality in a data science project
Figure 3: The reality in a data science project

Make data science work for your wealth management

To recognize and accept the inherent complexity of data science undertakings is an important foundation to get your use cases off the ground. Luckily, there are several steps that help a wealth manager overcome this complexity and shorten the resulting project cycles. In our report “Wealth management redefined using artificial intelligence” we listed 3 of these steps:

  1. Build a strategic AI roadmap
  2. Overcome integration complexity with a powerful service partner
  3. Profit from the deep learning effect

Read our report to get a set of 3 workable use cases, in which data is transformed into benefits for a financial organization and learn how to overcome data science challenges and circumvent common obstacles in wealth management.

Written by Gery Zollinger
Gery Zollinger leads the team behind the Avaloq Insight product line, which is designed to embed data analytics and artificial intelligence in wealth management. He joined Avaloq in February 2019 from Credit Suisse, in the global Credit Risk Analytics team, where he was responsible for credit risk modelling within the Private Banking and Investment Banking divisions. Gery Zollinger has worked in analytics and quantitative modelling for more than ten years. He holds a degree in economics & statistics from the University of Zurich (Switzerland), University of Lausanne (Switzerland), NHH Bergen (Norway) and in computer science from ETH Zurich (Switzerland).
Wealth management redefined using artificial intelligence

Download report

Learn more about:

  • The strains and prospects of a data science implementation project
  • How wealth managers create tangible business benefits with data science
  • 3 practical wealth management use cases of data-driven solutions