Asset managers see the potential for data science to generate alpha and increase sales but many struggle to achieve results. Investment models, alternative datasets, predictive lead generation analysis, marketing campaigns, client content, and more will continuously evolve as asset managers search for the winning or differentiated formula.
An enabling data platform is critical in a high volume, high velocity, and ever-changing data reality to help solve a manager's challenges:
Challenge 1: quant / data science data platforms must be able to quickly ingest data and provide the scalable processing power to find insights in new alternative and big data sources. Asset managers are adopting data lakes and cloud computing platforms to meet these needs.
Challenge 2: what is holding most firms back with their data science initiatives (and all initiatives) is providing easy access to high quality, operational data (holdings, performance, accounts, etc.). Data scientists spend too much of their time chasing and preparing data as opposed to what they were hired to do.
In addition, firms that are eager to advance their ambitions rush to sponsor disparate data projects by team, by asset class, and/or by distribution regions or channels. This fragmented approach results in the following issues:
- Redundancy: building multiple data platforms (clouds, data lakes, analysis tools, databases, etc.)
- Cost: paying for the same alternative or reference data multiple times
- Lost opportunity: not sharing insights, models, and solutions across teams
- Risk: creating data swamps (ungoverned data lakes), introducing InfoSec, and modelling risk
- Inefficiency: lacking the ease of access to high quality, operational data, and hard to ingest new datasets
Building the enterprise data platform and digital eco-system
Data foundation: providing ready-to-consume data (single sources of truth / MDM, organized data domains, data catalog, and managed data quality)
API catalog: to accelerate solution development, data access, and data discovery
Shared data lake: to quickly onboard and share new datasets (with the proper data governance and security)
Data aggregation tools: (ELT) to create unique datasets to feed data analysis solutions built in Python or R
Data visualization tools: to extract insights, publish results to dashboards, and inform decisions
AI/ML tools and data scientists will not produce consistent insights without the support of a modern data platform. The new data platform is not just for the data scientists and quants – all business functions should benefit from easy access to high quality data and the tools to garner insights. The benefits of becoming a data-driven asset manager will be realized when employees are empowered with a data platform built for the task.
In conclusion of Olmstead's 4-part Data-Driven Asset Management Insights series, we will discuss how to become a data-driven asset manager.
Contact the author: