Rage Frameworks Financial Statement Spreading How a Global Financial Services Company automated and scaled its Financial Statement Spreading operation, while cutting costs by 75% RAGE Frameworks, Inc. 2007-14. All rights reserved.
Introduction The general lack of capability to acquire high quality, reliable data from multiple and constantly changing sources in an automated fashion stands out in sharp contrast to the exponential progress that the rest of the business applications development world has made. This is despite the proliferation of file and document exchange protocols and standards, advanced OCR technologies, and markup languages. Business process applications are still designed to acquire data most effectively when the input is structured as per well-defined standards, which is often not the case even within the walls of a corporation, let alone outside. This shows up as a significant drop in end to end performance of otherwise well-tuned and automated business processes. Cost of the manual data extraction, rework, error handling and exception management, can easily represent 80% of the operating costs of an automated process. The reason why this area of such obvious waste does not make headlines, even as businesses look under every rock for an extra dollar to save, is that it has now come to be accepted as a necessary evil and as the cost of doing business. The Financial Services world, which faces these same challenges, has seen some positive traction over the past decade towards global standardization driven by XBRL.Org, an international consortium of firms developing XML standards to support the creation, extraction, and exchange of financial and business information, including the development of a taxonomy for financial reporting under US GAAP. In 2011 XBRL (extensible Business Reporting Language) was mandated by the SEC as the filing standard for all US public companies. However, during the period 2002 to 2012, the XBRL taxonomy went from being very rigid to virtually no standard structure at all in order to accommodate the unique reporting needs of the various constituencies. This underscores the continuing challenge of source data extraction in the financial world. The problems of normalization and data interpretation remain by and large unchanged. The extent of XBRL adoption beyond the scope of SEC mandates and whether or not and to what extent private companies and international companies adopt it are yet to be seen. In addition, paper documents and image files continue to exacerbate the issue. On the technology front, today s Application Development technologies are clearly a significant improvement over technologies five to ten years ago. In the data extraction space however, the solutions are grossly inadequate as the input document variability (form, format and quality) requires a degree of flexibility and rapid iteration that the software development lifecycle simply does not support. A new paradigm in place of standard SDLC is required. Manual approaches continue to dominate the market for data extraction and normalization. Outsourcing and offshoring are the norm to try and drive scale in order to keep up with business needs and advancement in downstream applications. This document describes how a Global Financial Services Company, a market leader in its domain, overcame these seemingly insurmountable challenges. Driven by internal compliance goals and external regulatory requirements, improving the quality and timeliness of risk reviews was getting increasing senior management attention. The document describes how the firm accomplished the following: Automated spreading Automated financial statement spreading, of over 100,000 statements annually RAGE Frameworks, Inc. 2007-14. All rights reserved. 2
Solution in 60 days Working spreading prototype for a set of public companies in 24 hours Enterprise scale solution in 60 days Controls & compliance Risk reviews based on reliable, ontime data Improved compliance Feeds to ratings agencies Spread data sent to Moody s Expansive coverage Domestic and international, public and private companies Deep cost reduction 75% cost reduction (for a process that was continuously tuned for greater efficiency ) Future ready The solution is fully automated, scalable, XBRL compatible (3 years prior to XBRL mandate) The solution can handle images, electronic feeds or paper documents Document Upload capability for secure upload of documents. The Business Problem As with their peer companies this firm found it increasingly difficult to manage the conflicting needs of lowering cost, managing growing volumes and quality of their spreading process, without any realistic possibility of automation now or into the foreseeable future. The firm s current technology did not allow for rapid extraction, interpretation, normalization and spreading of multi-formatmulti-source financial data without significant investments that would require replacing / upgrading functionality and would be disruptive beyond just this function and wasn t guaranteed to work. And again like their peers the company defaulted to a manual approach. And since managing credit risk is mission critical for the firm it continued to add people to address scale, improve data quality, better manage error levels, etc. Reorganizing for process effectiveness and fine tuning manual processes were not yielding results and not setting the firm up for the future. Frustrated by the sluggishness of manual extraction, manual spreading and associated higher risk exposure, and observing that XBRL wasn t going to be the silver bullet after all, the company s leadership asked the business and IT organizations to urgently come up with a long term solution. Challenges of Automation The key challenges the Financial Services company faced in automating the extraction, normalization and processing of financial statements were as follows: Source document variability: Financial Statements represent the core raw material in the production of the final product in the spreading process, be it credit rating or eventually a loan. Accurate and timely data extraction and processing of this data are critical. Financial statements come in Excel, Word, Quick Books and IRS tax returns. Formats varied, e.g. electronic Excel files, PDF, image (Tiff) files, faxed images or hard copy documents delivered via US mail. Together, the general variability of the source documents proved too much for the firm s conventional extraction approaches. RAGE Frameworks, Inc. 2007-14. All rights reserved. 3
Attempts at automation failed at the very beginning of the process. Statement variability: Each of the thousands of statements could come in a different format, often in multiple formats. Taxonomies used in these statements defining different transactions were not standardized, resulting in inconsistent labels for statement transactions. Analysts were often forced into clubbing data into the next most appropriate field, often without full understanding of the impact. The issue was amplified for international companies. Ensuring quality: Institutionalizing credit policy throughout the process and the organization was an ongoing challenge as the spreading teams (onshore and offshore) continued to expand and change through constant turn over. There were inevitable exceptions which required a skilled financial analyst, one with current policy at hand, to accurately handle. Internal / External systems: Access to the internal IT organization and the lead time for IT to build, test and deploy an interface to internal and external systems was unreasonably long even for a single interface. To develop multiple interfaces, and continuously keep them updated was not viable. Introducing a new company. Understanding and then manually extracting files from new companies was a slow and error prone process, that took several cycles to stabilize. Automated Spreading for Private & Public Companies Having scanned the market place for solutions the firm settled on a component of the Business Banking solution set from Rage Frameworks, driven by the following considerations: 1. Rage was a technology platform that was purpose built, ground up, on the principles of rapid implementation, flexibility and scalability. This was demonstrated via a working prototype delivered in 24 hours with live spreads of financial statements of several companies. This was well beyond the firm s expectations. 2. Ability to service Enterprise scale operations while being agile and flexible, both with the core platform and the partnership. Rage had developed a very strong (and patented) set of extraction components which had been stress tested across millions of transactions, on a daily basis for over a decade. 3. Deep financial services domain expertise. Financial Services is Rage s area of focus. Most of its business today continues to be in this sector. Rage combined its extensive experience with its patented extraction technology to extract financial information from any source (feeds, uploads and paper). The solution was always live, capable of being modified on the fly, rapidly and without any maintenance overhead. Starting with Rage s out-of-the-box client portal, configuring the system for the specifics of the firm s spreading policy and integration with internal and external systems, an enterprise-scale solution was operational in less than two months, addressing all of the following fundamental automation challenges: Document type variability: All documents continued to be sent by the exact same method and in the same format. Electronic feeds and uploads were immediately available for the extraction process. Physically mailed documents were scanned using Rage s OCR wrapper and content extracted, quality checked and made available for normalization. RAGE Frameworks, Inc. 2007-14. All rights reserved. 4
Statement variability: Real time modification of the business logic made dealing with financial statement variability straightforward, as demonstrated time and again during the implementation. No code was generated; the business user drove the process. Quality: By incorporating consistent business logic throughout the application, standards and credit policies were institutionalized. Oneoffs were also handled as per policy. Internal and External systems: Integration with internal and external systems occurred faster than ever before as protocols and mapping were handled and configured as data to create batch and real-time interfaces. Implementing new statement formats: Using metadata driven mapping the Rage Platform makes this simple, something that can be put into production in a few hours. The solution was tuned interactively until the spreads were fully automated (backed by an exception management team). Even poor quality scanned images sent over by the clients were handled mostly in an automated fashion. Starting with a handful of companies this was rapidly scaled as the ease and effectiveness of the solution became apparent. Expanding the Solution Today the firm processes 100,000 spreads in 46,000 public and private companies in North America, South America, Europe, and Asia pacific, in multiple languages. Financial statements in all formats, including XBRL are uploaded. Hard copy source documents are imaged and tagged and are made available for review and audit purposes. Rage differentiating feature: Aggregation across multiple sources, including XBRL feeds, reporting mediums [web scraping, mail, feeds, excel, images.] built on patented technology to extract data from financial statements. The As reported information is extracted and loaded into a repository and made available for review and analysis via a portal. Automated rules based quality assurance steps are embedded throughout the process (e.g. ensuring that the balance sheet balances at the end of the extraction stage). This drives accuracy at each step and provides traceability for troubleshooting where required. Exceptions are handled through the application portal by highly skilled Rage analysts and client analysts, as required. Secure, self-service document upload capability Portal for comprehensive visibility of source and spread data, for effective review, audit trail and exception Normalization is driven by rules specified by the client, IFRS, XBRL, etc. Foreign Financial statements are interpreted and non US GAAP statements are normalized to US GAAP. Unique features and differentiators of the Rage Spreading solution: Document source and format agnostic. Secure document upload facility for statements Industry specific normalization of data Analysis of revolving credit lines Auditors opinion on the financial statements is captured Key break-ups from notes to financials Industry Identification data: NAICS, SIC or GICS codes General Corporate, Financial Institutions, Muni, NPO, ITR RAGE Frameworks, Inc. 2007-14. All rights reserved. 5
Calculations for extraordinary/one-time/noncash items Extraction of details on operating leases and other contractual obligations Automated Quality Assurance checks Handling of foreign languages statements Key actions in each process step: Document upload Self service, secure documet upload facility (source document and format agnostic) Extraction Process Extract tabular and non-tabular data from documents Embed anchors into the document for click-back from final views Apply longitudinal historical data to current extraction to interpret extracted line items to normalized templates Discover Sub totals Computational Process Calculate ratios and value added data Provide transformation of data where required Product Process Create "As Presented", "As Reported" and "Normalized" products Propagate extraction links to product views Workflow Process Queue management, action Item assignment and alerts. QA Processes (validate extraction completion, sub-total checks, generation of normalized statements, client specific checks) This capability has now been extended to other business units of the company with different business systems architectures, seamlessly and rapidly. Enabled by visionary leadership, a market leading process-technology platform, a strong partnership with the solution provider, the global Financial Services firm has a solution that meets all its current needs and is future ready. The entire solution was implemented using patented Rage technology to create an extensible, scalable and highly flexible platform with potential application significantly beyond the current scope (e.g. adjacent features, such as scoring, rating, bureau access, rule-based generation of intelligent, analytical reports, risk based pricing) are built in to the deployed solution and can be easily turned on and configured. Rage Platform, designed and built for flexibility: Insulation from frequently changing source file formats. Ability to add a new format rapidly. Changes in nomenclature, e.g. the emergence of new IFRS, XBRL guidelines; evolution from an extraction and normalization perspective. Ability to integrate new companies RAGE Frameworks, Inc. 2007-14. All rights reserved. 6