The way forward for portfolio management? Hyper-personalization!

Individualized customization tailored to a consumer’s exact wants and needs are just some of the ways hyper-personalization is manifesting in today’s world. The financial industry is no stranger to such a trend, but how can financial advisors keep up with the sheer amount of customization clients are demanding?

Hyper-personalization is the next big thing in business. Nowadays, companies are competing for the chance to prove that they “know you” best, whether it be in the use of predictive algorithms to recommend new products, videos and TV shows (Amazon and Netflix literally have this down to a science), or by presenting usage behavior in pretty charts and exciting statistics (such as in Spotify’s “Yearly Wrap”). With every industry pushing to increase personalization and curated experiences, the financial industry needs to do the same.

In this blog we’ll focus on how this plays out for portfolio management.

The challenges of portfolio optimization

Portfolio construction and optimization are often the most important and difficult tasks for financial advisors due to the wide variety of different, and sometimes contradictory, factors they must consider. For example, diversification is recognized as being beneficial in portfolios, and often the more diverse the better. However, clients often have other factors that they consider to be important (like a regional preference due to a home bias), even when such preferences contradict the idea of diversification. In addition, a particular investor may find other factors important (and what’s important to one investor differs from what’s important to another). Examples of these include Environmental, Social and Governance (ESG) considerations, asset classes weighting, ETF exposure, a focus on specific investment themes such as mega trends, dividend payments and more.

In the age of personalization, clients expect to be able to fully customize and outfit their portfolio according to their unique wants and needs. For advisors this makes finding the optimal portfolio a complex and time-consuming process, one that requires a deep knowledge of financial markets as well as a strong understanding of documentation and regulatory procedures (not to mention knowing the client and his or her particular objectives). In short, there are simply too many factors for one person to handle all by themselves, and not enough information to do so using traditional approaches.

Traditional methods can’t keep up

In today’s world traditional methods – such as the MPT Model developed by Markowitz in the 1950s involving the importance of portfolios, risk, diversification, and the connections between different kinds of securities – as well as other legacy models – don't work optimally because they can't deal with the complexity of so many factors. Markowitz portfolio optimization, for example, is a one-step procedure that simply maximizes the portfolio return subject to a given risk constraint.  Furthermore, the results of Markowitz optimization are very sensitive to the input data. Therefore, more robust solutions that can handle multiple factors are needed.

The answer lies in Artificial Intelligence (AI)-based genetic optimization

This is where technological advancements will help, specifically AI-based genetic optimization. This optimizer looks at a portfolio like a chromosome composed of a personalized genetic mapping. Using Darwin’s theory of evolution, which is based on the principle of natural selection, portfolios are being slightly modified, retested and ranked according to their fitness – their ability to match the client’s preferences. Portfolios with poor fitness are discarded, the best form the next generation. This computational process based on algorithms enables artificial learning; the optimizer learns with each generation until the optimal portfolio is found.

Algorithms based on the principle of Darwin's selection theory are already widely used across various industries. They form the basis of Google AI, are used to calculate the airfoil profile of aircraft and can even plan robot movements. In AI-based portfolio optimization, the system may at first generate 50,000 different portfolios, purely randomly, measure the factors of each, weigh the factors and calculate the overall fitness score for all 50,000 portfolios. To get to the end-goal faster, portfolios with high fitness have a higher probability of surviving and portfolios with low fitness have a high probability of being removed from the genetic code or optimization process.

Moreover, portfolios with very good fitness have a higher chance of reproducing and continuing onto the next generation. Combining the best characteristics or investments creates better portfolios. In technical jargon, this process is called crossover and it can be repeated numerous, often hundreds of times. After a certain number of generations, the overall fitness score will approach the maximum fitness score and you end up with the best possible portfolio that is individually tailored for the client.

How long does this process take?

A major benefit of the AI-based genetic optimization process is that it can be completed within seconds. Clients can also modify and edit the portfolio in real-time, as well as get a complete overview of their investments with the help of the AI-generated charts and statistics. Furthermore, the number and type of fitness factors is up to the client and can be in constant development and improvement. There are several types of standard factors – financial figures such as return, volatility, Sharpe ratio or currency exposure, along with factors like sustainability, ETF, asset class or regional exposure –  but the entire build is your client’s sandbox.

AI-based optimization improves the quality of life for advisors and their clients and makes complex financial decisions much easier. It’s not a stretch to say that this is the future for portfolio management. While the personal and human aspect of trusted advisors remains essential, the underlying manual calculations and time-consuming efforts can and should be replaced to meet the varying needs of each and every client. Hyper-personalization via AI is the way forward.