Structured and unstructured data
Structured data is data that has been organized into a formatted repository, often a database, so that the information can be made accessible for more effective processing and analysis. This type of data is typically numeric such as performance and benchmark returns, and portfolio characteristics, like P/E and Sharpe ratios.
… taking advantage of the next technology revolution and the benefits of automation will be much easier if firms have well-defined data that’s easily accessible. Cindy Sealey, CFA, Principal – Practice Lead, Cutter Associates
Unstructured data is data that is typically not contained in an organized database; the data is usually in the form of narrative, images, or even video. Examples of unstructured data are investment commentary, footnotes, and portfolio manager biographies. Historically, this unstructured data was created in near identical iterations each time it was used and resided within one specific document rather than in a centralized location for reuse.
As this unstructured data has not typically been centralized, chances for inconsistencies increase. Client Reporting, Marketing, and Sales Staff must wade through emails or other documents to find the most current version of the text. If the same unstructured data is presented in multiple materials, any changes to that data need to be propagated across all, but getting unstructured data changed throughout these materials can be a challenge.
Adding structure to unstructured data
But what if you could re-use unstructured data the way you re-use structured data? The key is identifying the unstructured data elements used across the firm and organizing them. That’s a huge undertaking, but some investment managers are starting to recognize the benefit potential for having defined data elements and are reconfiguring how they organize and store them. (see Figure 1).
How does your firm store unstructured client data?
Source: CutterBenchmarking, Client Reporting Survey 2016
Think of an investment commentary as an example. That one data element could appear in a client report, a client account review, a pitch book, a web portal, etc. How do you successfully manage a change to that data across all those platforms?
To ensure consistency, reduce risk and duplication of effort, and automate workflows and review processes, firms need to add structure to their unstructured data. They do this with centralizing storage, applying appropriate tagging and metadata, and by building out templates that allow for automation of this data into the templates for material production, with each element such as investment philosophy, investment process, footnotes, graphics, logos, etc. making up the whole.
Understanding each data element’s meaning and its relationship to other data elements is critical. Establish a common set of data elements for unstructured data topics – for example, a Portfolio Manager’s biography may consist of First Name, Last Name, Suffix, Industry Designations, Title, Short Biographical Text, and Long Biographical Text.
Role for data governance
The first step in any attempt to efficiently convert the way a firm creates, stores, and utilizes unstructured data is to make it part of the firm’s Data Governance program – i.e. the processes that ensure that important data assets are formally managed throughout the enterprise. If unstructured data is related to structured data, a firm should use the existing group who governs that data domain. For unstructured data without a data domain established, a firm should include representatives from areas responsible for its creation and usage. Consider investment process text as an example. Representatives involved in its governance process might include Portfolio Managers, Product Management, Sales, and Marketing.
Second, you’ll want to understand the current workflows around the data – Who creates it? Who approves it? Where it is stored? How often is it updated or reviewed? Who uses it? Any existing processes and procedures should be examined for redundancies and opportunities for streamlining and improvement.
Third, you’ll want to create well-defined roles around the data – creator, owner, steward, IT custodian – recognizing that one person may have more than one role. These will be the people to hold accountable for making sure the data is correct and ready to be used.
Creating a storage place
With data governance in place, firms can determine how best to store their unstructured data. A variety of options exist today including NoSQL databases, Content Management Systems, Document Repositories, and Data Lakes. Each of these allow for a storage location which can feed downstream systems used to automate the production of materials such as client reporting systems, presentation tools, RFP tools, consultant database population tools. Data can be created once, approved by compliance for specific uses, automated to the material templates, and have event-based triggers such as expiration and review.
With the data well-defined and organized, it is possible to integrate it into solutions that allow pitch books, client account reviews, and factsheets to be created at the touch of button. Those responsible for unstructured data creation may also be able to see which data is being used through tracking metrics and eliminate data which is not. Data can be adjusted to be more impactful where needed, and Marketing can quickly create and disseminate content that resonates with prospects and clients. Digital distribution will be easier as both client and internal demand for interactive tools and online information increases.
With the data well-defined and organized, it is possible to integrate it into solutions that allow pitch books, client account reviews, and factsheets to be created at the touch of button. Cindy Sealey, CFA, Principal – Practice Lead, Cutter Associates
Other impacts to delivery
Firms are facing growing client demands for more timely information, accurate data, and relevant insight which often take the form of reporting. But what if your firm could provide insights more quickly? Commentary and any other written narrative in client reports tends to cause delays to the delivery of reports after period end. Despite improvements in the speed of closing an account’s books or calculating performance at a period end, the human aspect of writing commentary can be a bottleneck to report distribution. Almost 46% of Cutter members take 10 business days or more to deliver client reports. (see Figure 2).
How many business days to produce your client reports?
Source: CutterBenchmarking, Client Reporting Survey 2016
So what are firms considering to speed up this function? Some are deploying natural language processing (NLP) tools to “write” commentary using input from performance returns and performance attribution calculations. While this is easier to do for passive strategies, Cutter Associates expects with artificial intelligence these tools will be able to learn the algorithms humans are deploying and evolve for further uses.
Need for continuous development
Competition in the investment industry is fiercer than ever. Firms recognize that client and prospect engagement is critical – and that the data provided to these constituents needs to be accurate, timely, and appropriate. No firm wants their client facing tools to be the equivalent of Nokia or Blackberry as their competition has moved on to Samsung or Apple, so it is important to keep pace with technology advances and be ready to experiment with new tools. Managers who are willing to adapt and invest in new technology to automate the critical tasks that support distribution will likely find themselves ahead of the competition.
About the author
Cindy Sealey, CFA, Principal – Practice Lead, Cutter Associates
Cindy Sealey has over 28 years of experience in the investment management industry. She leads Cutter Associates’ Client Facing practice. She consults to investment firms on their institutional sales & client service practices, operations, data management, and strategic technology direction. Cindy has conducted system searches for client reporting, RFP, CRM, performance, accounting, performance, portfolio management, risk management, equity research, and GIPS composite systems. Prior to joining Cutter Associates, Cindy served as Vice President and Head of Operations for a division of Guggenheim Investments, where she led mutual fund accounting and administration; institutional accounting, billing, and client reporting; SMA operations, performance reporting (GIPS®); fixed income performance, attribution and risk reporting; fixed income investment team support, insurance reporting, asset management technology, and equity and fixed income trade settlement. Cindy holds a B.B.A. from Washburn University and is a Chartered Financial Analyst.