Michael Shashoua: Protecting Valuable Data

A recent controversial psychology experiment conducted by Facebook reportedly tested whether skewing its users’ news feeds toward positive or negative stories would affect the nature of the posts they then made. However intrusive this appears, it’s actually not that different from financial industry data analysis.

Based on the securities data that firms collect, data analysis is used to determine whether to react positively or negatively—to buy or sell, essentially. Although, as Toronto-based data scientist Hashmat Rohian of Aviva Canada, says, financial services firms would attract legal and regulatory scrutiny if they conducted an experiment like Facebook did without clients’ consent, even though data collection by itself is a regular part of operations.

With the high volume and value of data currently being collected, firms must be concerned about protecting the integrity of that data and meeting privacy requirements, as well as cost and risk issues.

Even without a conscious overreach, like Facebook’s, breaches of data security often turn up in the news. If more data is collected, linked, connected and cross-referenced, then it is all the more important to securely store and protect the resulting analysis and insights. In developing or collecting ever-more-complex types of data that contribute to more insightful analysis, whether for risk management or to support trading desks, firms are generating a valuable resource and should appreciate its value and business potential.

With these concerns in mind, and with cloud computing providers trying to make inroads into financial data management, broker-dealers, buy-side firms and exchanges have expressed support for standards that are consistent, yet open, and for protecting private information, according to studies conducted by eTrading Software, with Boston Consulting Group unit Expand, and by Aite Group.

“We recommend a combination of a taxonomy, a data model and a transport layer,” says Alex Wolcough, head of collaboration practice at eTrading Software in London. “Collectively, that will create a mechanism that will work for the market as a whole.”

Markets with strict privacy rules are slower to embrace cloud for data storage and operations, notes Virginie O’Shea, senior analyst at Aite Group and a co-author of its study. “Privacy is an issue if you have something commercially sensitive,” she says. “You may want to have this built into your agreements.”

Even where the regulatory environment is more permissive, it is advisable to mitigate operational risk to ensure data remains available for auditing, and to meet transparency demands.

Searching For Silver Linings
Technology transformations involved in going to cloud resources are also a concern for those responsible for data. You can’t protect and secure data if it is lost or corrupted due to faulty systems, or a lack of operational risk mitigation.

So far, the industry has favored private clouds, as Aite Group’s study notes. But while private cloud computing is more secure and can be dedicated full-time to an organization’s specific needs, public cloud resources are less costly, which is a good attribute when resources and budgets are a concern.

Keeping the value of data in mind when deciding how to manage it, one must be wary of making the penny-wise choice that could prove foolish if it allows security breaches, analytical errors or additional operational risk.

The industry, once reluctant to use any kind of cloud at all, is beginning to explore private clouds, with tier-one firms building their own private cloud resources. “There hasn’t been massive take-up yet for these utilities, not even on the vendor side,” O’Shea says. “But the potential is you could put all the data up there in the future if you’re using a private cloud.”

Cloud’s slow adoption may reflect mindfulness of privacy, security and cost issues. These shouldn’t preclude use of cloud computing completely, but ought to guide how firms proceed.