Large asset management firms like TIAA-CREF produce and collect a great deal of data. But are they collecting it, using it and storing it in the most efficient and cost effective manner?
With nearly 4 million participants, more than $500 billion in AUM and representing more than 15,000 institutions, asset manager TIAA-CREF has more than its share of participant data.
The firm is just one example of the myriad asset managers and institutional investors seeking data consolidation and cost containment solutions.
At the SourceMedia MDM & Data Governance Summit in New York, TIAA-CREF, along with Princeton, N.J.,-based Global IDs, a technology firm that provides companies methods to manage their information, discussed the challenges of the big data problem and finding answers.
"Our goal," says Arka Mukherjee founder of Global IDs, "is to take costs out of these very large data ecosystems. And when we talk about a financial services company or large pension fund, there is a tremendous amount of complexity in their data ecosystems. All that complexity does is add to costs - and the worst part is, the costs keep going up."
The questions that arise for TIAA-CREF, as a result of increased participant data, company officials said, are how best to formalize data governance, create mature data management practices, optimize it and enable multiple businesses to access it. It's not a simple task and it takes time, effort and money to create a system where perhaps there wasn't any before, or one, which was simply not keeping up with increased needs.
To get a handle on its big data, TIAA-CREF presented several goals:
* Data Governance: formalize business data owners and data steward roles by developing a playbook.
* Enable Business Initiatives: linking data governance and stewardship to key programs.
* Build Enterprise "Big E" Data Platforms: creating horizontal data platforms that enable multiple strategy and business initiatives.
* Optimize Data Infrastructure: consolidating or retiring databases, modernize technology and automate enterprise data platform administration.
* Transform Data Management Capabilities: create enterprise data services for broad use.
DATA GOALS
But getting to those goals takes time and effort. Mukherjee says four things must occur before a large financial services firm can begin changing the way it manages data.
First, he says, firms need to understand their data assets and take a complete inventory from top to bottom -- something called semantic profiling. The second issue is to address any problems with the way governance or lack of governance of the data assets is being conducted. Third is “rationalizing the assets” meaning taking out duplicates or “low value” assets (data that is no longer relevant or old, for example) from the data. Finally, notes Mukherjee, is the need to reduce costs by implementing a cloud service so that large, expensive consultants and software can be eliminated from the equation.
Paul McInnis is the head of enterprise data management with Eagle Investment Systems in Wellesley, Mass. He says a challenge facing large firms is determining what their “big data” strategy is and in doing so considering the differences between unstructured and structured data and all of the various mediums within which to store the information.
Unstructured data, for example, would be taking data from call centers or emails, or even news reports and seeing how that information can be analyzed. Structured data, on the other hand, is more tangible (less anecdotal) information, from transactions, subscriptions and distributions (or members, as in the case of a large asset manager or pension fund).
“It’s early days for investment firms but it definitely has everyone’s attention,” notes McInnis. “From my experience, our clients are looking at structured data and the analytics that come out of it.” Practically speaking that could mean using the data to assess and compare the quality of vendor information, identify trending statistics by asset class or evaluating their overall operational efficiency, for example.
Much like Mukherjee has stated, McInnis says it is important to weed out the bad data and remove the “false positives” from the equation. “Quality is the biggest challenge when you are looking at this because the whole purpose of this is to make decisions off of that data. So you need to know that data is good.”
ENSURING QUALITY
Another important point is the ability to compare apples to apples when it comes to data. Companies, notes McInnis, will develop their own glossaries in order to do so.
In the case of TIAA-CREF, a “playbook” was developed that helps define, control, monitor, expand and optimize data so that quality is ensured and shared services can be optimized and supported.
In other words, notes McInnis, lineage allows users to track consistency of data from “integration through consumption” while allowing users to know the quality is consistent across multiple sources. Some firms also have so much data that some of it must be archived or is no longer relevant. However, the lineage is important to ensure consistency and that data quality permeates across multiple sources over a long period of time.
And from a data governance perspective, companies will sometimes realize they have various definitions for the same term or value. “Storing that data on one platform drastically increases the quality and the confidence you can have when you are analyzing that data,” explains McInnis.
McInnis and Mukherjee are in agreement in that data must be looked at as a whole, from one reliable source, and not in silos.
On a simpler level, Mukherjee says low quality data can be frustrating to the user and the customer if erroneous information is being inputted or out-of-date data is still part of the database. Similarly, he notes, asking multiple databases to provide the same information on a particular topic might garner different responses if the input information is faulty.
Overall, notes McInnis, companies need to view their information as an asset. “As an asset it needs to be managed as a resource with value and I think that’s what firms are realizing today.”
