The Chief Data Officer MG Marcan B.Sc Eng (Wits), M.Eng (Wits)

Abstract – The role of a chief data officer (CDO) entails the accountability over data creation, usage and management. Due to duplication, legacy systems and organization structural changes – data quality continuously drops and must be centrally managed to reduce the risks associated with poor data quality. Governance, data stewardship1 and data ownership can help facilitate the management of data and its quality.

I. INTRODUCTION This paper defines the Chief Data Officer (CDO) role, details the associated areas of responsibilities, and outlines key strategies for the role. Readers can use this paper to understand and benchmark this role, or use it as a guide to formulate and define a CDO role within their organization. II. BACKGROUND A. Definition The Chief Data Officer is an executive management role for enterprise-wide data processing and data mining [1]. There are other names for this role, such as: Global Head of Data [2] and Chief Information Quality Officer (CIQO) [3]. The role requires business, technology and diplomatic skills, and must be given the authority to make decisions that have an organization-wide impact [4]. B. The Problem Data is a strategic asset, and is often a direct or indirect source of revenue. It is both extremely important and underutilized in many areas [5]. The importance of the accuracy and reliability of data can be easily demonstrated. For instance in financial trading, if reference data (such as instruments, rates and counterparty) is sub-standard, poor trading decisions may result, which can be costly for the business [6]. Without a holistic and managed data capturing processes, issues such as sub-standard reference data, rates mismatches during finance reconciliation, and difficulties in tracing capital spending (e.g. matching exposures and hedges) - can arise.

very difficult to use to generate a holistic view on the data within the organization [7]. Furthermore, having data silos mean separate strategies and environments; fragmentation of infrastructure; loss of staff synergy and loss of technology and information sourcing opportunities (reuse). In addition, increased complexity and risks can arise which have negative compliance, competitiveness and cost implications [6]. Some statistics demonstrating the consequences of poor data quality include: 10-25% of the total revenue loss of any organization can be attributed to poor data; the top technology reason for CRM failure is poor Data; the Data Warehousing Institute estimates US business Data Quality problems cost more than $600 billion per annum [3]. C. The Solution The actual source for disjoint data is within the process of how the information is obtained, stored, maintained and distributed. This is also where the solution must originate. In addition, a centralized drive for managing the quality of information should also be followed [7]. Clean, qualitative data, strengthens client relationships and reduces risk. By improving data creation (through removing redundant and conflicting data), risk and control can be improved, data will need less repairs and the downstream data will become more efficient [2]. Each organization has its own priorities and challenges, and opportunities for improved quality and service have been noted to be the hallmark of data management over the foreseeable future [8]. Centralizing data enables more creativity and more cost-effective vendors with more functionality. Synchronized data ensures consistency, completeness, reliability and high quality of results [6].

Sub-standard data quality can occur due to duplicate data repositories, the existence of legacy systems, internal changes and the after effects of external acquisitions. An additional contributor is the fact that either too many people are attempting independently to solve a data problem (creating unconnected solutions) or no one is. The disjointed pockets of information are

D. Is this new ? The role of the CDO was not notable prior to the 1980’s and has emerged from the fundamental functioning of everyday business (information technology, Business Intelligence (BI), data integration and data processing) [1]. Today we see several examples of a CDOs such as: Dr. Usama Fayyad (Yahoo); Thomas Mueller (Allied Management Group - AMG-SIU); John Bottega (Citi Markets & Banking) [8]; Peter Serenita (JPMorgan) [2]; Michael Spears (NCCI Holdings) [9].

1 the formalization of accountability for the management of the data assets. collecting data, converting and maintenance; data management standards, procedures, and accountability; governance focus, policies and guidelines for data management processes [4]

Changes in the international data quality landscape are also apparent: the Swift introduction of International February 2008

Bank Account Number (IBAN); Kaupthing bank (Icland) implemented a central market data pool for the whole bank; Citi has introduced a first release of the data standards hub—which distributes standards information, definitions and structures [6].

B. Evaluate •

III. RESPONSIBILITIES The CDO normally reports to the Chief Technology Officer (CTO), the Chief Executive Officer (CEO) or the COO (Chief Operations Officer) [6]. In the financial world, the role extends from financial instrument repositories to customer and account data, corporate actions, pricing, positions and transactions [7]. It includes front and back office activities as well as the integration and the data flow between them, operation, finance, compliance and all risk areas [6].

•

• The CDO must: A. Drive Governance Drive the processes of creating data [2]; derive best practices [7]; create and manage data standards across the entire organization [2]; drive policies and procedures relating to data governance, data ownership, data sharing, and data privacy [4]. B. Manage All data must be managed in a centralized manner [7] as a critical discipline within the organization [6]. The CDO must place strategic priorities on data systems and related opportunities; extend existing data services outside their normal operational area (to reduce redundancies) [2]; optimize revenue and business generated via data and its management; and recommend best strategy for acquisitions (e.g. integration and use of global data standards). C. Lead The CDO must drive data as an revenue generating asset to all business stakeholders (executives, employees, customers) [2]; extend data quality and prioritization, provide support to research and analysis, trade execution, data matching, settlements, reporting, valuation, risk, and compliance [6].

C. Centralize •

•

•

•

•

A. Create Focus on

• • • • •

Data and the processes around the creation, distribution and usage of data. Prioritizing data in business requirements, and driving technology as the implementation [6]. Driving data governance: data structure, data ownership, stewardship, and operations [6]. Driving data policies and procedures that govern the acquisition and use of data. Ensuring a strong senior management governance structure [6]. Ensuring data models and processes are simplified and enforced [3].

Collaborate in industry-wide data management efforts. This is particularly important in the finance industry, as data metrics, at the moment, are nearly nonexistent (which makes it difficult to benchmark) [8]. Carefully consider issues such as flexibility and responsiveness verses enterprise-wide data management strategies; Market data verses Reference data management: recognizing similarities and differences. [6] Manage data distribution: Creation and standardization of data models, establishment of a common ground for data communication [8]. Create both Reactive and Proactive Components [3].

D. Coordinate

IV. CDO STRATEGIES •

Create and / or utilize metrics and other measures to assess the delivery of data management. Examples include: the Information Maturity Levels (IML) ; the Enterprise Information Quality Quotient [3]; and creating metrics based on end to end subset of functions across slices of the business processes [8]. Create and /or utilize a maturity model to assess and drive the level of data as a strategic asset. Example includes: The Information Quality Management Maturity (IQMM) which includes 5 levels: Reactive (BI, Data Warehousing and uncertainty); Awakening; Enlightenment; Wisdom and Certainty [4]. Benchmark and identify benefits: For example, based on TDWI Data Quality Survey, major benefits include: Greater confidence in analytic systems, less time spent on reconciliation, single version, and reduced costs, (March 2006, 750 respondents) [3].

•

• • • • • •

Manage Data Suppliers: selection of methodologies, understanding expectations, performance, and providing quality assurance [4]. coordinate data management efforts: facilitate between the business, technology and operations sides to implement proper data management solutions [6]. Prioritize data cleansing of existing data [8]. Work towards a “Golden copy” for data for descriptive and reference purposes. Utilize external data models such ISO and other widely accepted conventions [6]. Leverage off the firm’s best subject matter experts across the various business units[7]. Help make data actionable by business [4]. Work closely with compliance [3].

February 2008

E. Prioritize • •

• •

Clean up data to satisfy regulatory issues [2]. Rationalize the current environment: Inventory all of data systems and bring the environment down to a manageable number of systems [2]. Enforce data policies compliance [7]. Encourage data-aware culture to extend thinking beyond immediate areas of responsibility [2].

F. Manage •

• • •

•

•

Arrange methods and standards for the distribution of data to down-stream consumers [6]. Drive data stewardship. Split data stream into different sub-streams based on how it needs to be treated. Ensure data privacy and security policies, whilst preserving and driving the data's statistical value to make predictions and facilitate real-time analysis [5]. Engage and drive solutions [7]: in-house, external packages, processes revamp and automation [6]. Make data Transparent and Traceable: Have clear and documented definitions and procedures and well-assigned responsibilities on data; Manage well-structured access and security of data [10]. V. CONCLUSION

The CDO is an important position within the organization responsible for management, coordination, prioritization and quality assurance of data. The primary focus of the CDO is to define, drive, manage and account for the creation and usage of data within the organization. There are various strategies described in this article which demonstrates how these responsibilities may be approached. Nonetheless, the priority, structure and tactics to implement these strategies remain the challenge of every organization independently.

[5] Boyne J.,Master of the Data, http://www. intelligententerprise.com/showArticle.jhtml?articleI D=59301282, Published 1 March 2005. [6] Inside Market Data, Reference Data – Special Report: Make your data integrated, http://db .riskwaters.com/global/IMD/pdfs/Oct07_IMDRe ferenceData_web.pdf, Last modified 24 October 2007. [7] CXO, Hot Data, http://www.cxo.eu.com/pastissue/ article.asp?art=271470&issue=220, Last accessed 4 February 2008. [8] Wakem M., Data Remediation, Rationalization and Distribution Top CDO’s agenda, http://www .wallstreetandtech.com/features/showArticle.jhtml? articleID=201201425, Published 26 July 2007. [9] Mac Sweeny G., Citigroup Names Bottega CDO, http://www.wallstreetandtech.com/showArticle.jht ml?articleID=183700099, Published 16 March 2006. [10] Business Insights, The Future of Regulatory Compliance, http://www.globalbusinessinsights .com/content/rbfs0067m.pdf, Last modified 24 October 2007. M. Gabriel Marcan: (Born 1976) Masters in Engineering (Cum Laude) (Electrical and Information Engineering), University of the Witwatersrand (2005), South-Africa. B.Sc Engineering (Cum Laude) (Electrical and Information Engineering), University of the Witwatersrand (2004), South-Africa. He is currently employed as an APPLICATION SPECIALIST with Corporate Investment Banking (CIB) Global Markets Technology at Standard Bank (South-Africa). Previous work includes a SOFTWARE ARCHITECTURE AND DEVELOPEMENT with Nyper Networks (South Africa). He is interested in information and software engineering, information processing and management, intelligent search engines, artificial intelligence, communication and physics. Mr. Marcan has won several awards during his academic studies including: IBM - Best Student Recognition Award (JHB 2005), Post Graduate Merit Award (Wits 2005), Merit Certificate (Measurement Systems) (Wits 2004), E J A Loerincz Scholarship (Wits 2004), University Council Merit Scholarship (Wits 2003), Golden Key International Honour Society member (Wits 2003), Merit certificate Electronics(Wits 2002), and Merit certificate - Physics (Wits 2002).

VI. REFERENCES [1] wikipedia.org, Chief Data Officer, http://en .wikipedia.org/wiki/Chief_data_officer , Last Accessed 4 February 2008. [2] Valentine L., Citi’s Chief Data Officer Turns Data Into a Business Asset, http://www.financetech.com/ featured/showArticle.jhtml?articleID=201307805, Published 1 August 2007. [3] Malakar D., Principles of ROI Development for Data Quality Projects, http://www.tdan.com/viewarticles/6701, Published 1 January 2008. [4] Malakar D., Chief Data Steward or Chief Data Officer: Another C-Level Acronym?, http://www .tdan.com/view-articles/4581, 1 January 2007.

February 2008