How to Deliver Enterprise Data Quality Management: The Jay Zaidi Interview

Jay Zaidi is Enterprise Information Management Leader based in Washington D.C, USA

Jay Zaidi is Enterprise Information Management Leader based in Washington D.C, USA

How to Deliver Enterprise Data Quality Management:

The Jay Zaidi Interview

What does it take to deliver an holistic enterprise data quality solution across one of the world’s largest and most data intensive organisations?

In this detailed interview, Jay Zaidi provides a huge amount of practical, actionable advice for anyone looking to take their data quality initiative to the wider organisation. 

Sample of topics include: the importance of collaboration for data governance and data quality management, evaluating COTS products for Enterprise Data Quality initiatives, Jay’s career path in the profession, integration challenges on large enterprise data quality projects and what stumbling blocks to avoid on complex initiatives.

Image credits – Creative Commons VinothChandar

Data Quality Pro: For the benefit of our readers can you briefly describe your current role?

Jay Zaidi: I currently lead the Enterprise Data Quality (EDQ) Program at a large Financial Services firm. In this role, my primary responsibilities are to deploy an EDQ Platform, EDQ Services, and an EDQ Issues Management system and provide consulting services to all lines of business (LOB). These foundational components are utilized by LOB teams and the EDQ team to monitor, measure and proactively manage the quality of business critical data across the enterprise. My team works closely with the Enterprise Data Governance team and the various Data Custodians and Trustees, to ensure alignment and compliance with the Enterprise Data Management standards and policies.

Another critical aspect of the team’s role is to develop data quality best practices, data quality development frameworks for the enterprise tool set, configuration management guides, design patterns and guidelines documentation, templates for various Software Development Life Cycle (SDLC) artifacts, and training material. We also consult with the LOB’s and Senior Management on data quality and data integration projects and build enterprise data quality web services, which can be leveraged by various applications. 

Data Quality Pro: At the onset, why don’t you tell us what the key drivers are for firms to consider implementing data quality across the enterprise? Why invest in such an initiative? What’s the business case?

Jay Zaidi: Data is a company’s lifeblood. Businesses run on data – data plays a very important role in corporate decision making, financial reporting and disclosures rely on data, and risk is monitored and managed using data. Therefore, it is clear that data is a company’s strategic asset and should be treated as such. The questions we should be asking are – Do we treat data as a strategic asset, and are we making the right investments with respect to defining data quality requirements, proactively monitoring the quality of business critical data and managing it’s quality throughout the business value chain. 

New legislation and oversight bodies are being instituted to monitor and address systemic issues related to data quality, particularly with multi-nationals and government regulated firms. Large firms in the Financial Services sector have noticed this and are taking steps to implement holistic data quality controls. Such issues are not exclusive to firms in the Financial Services sector, but apply to all business sectors such as Health Care, Pharmaceuticals, Consumer Packaged Goods, Retail, Energy, Manufacturing, Government, Transportation, Real Estate, Electronics, etc.

This is a Board and C-level issue. To be frank, ensuring that the consumers of corporate data have the highest level of confidence in it, is not just good business practice, but is an imperative for companies that wish to prosper in the 21st century. Each firm should consider the significant human and capital resources it expends in addressing data quality-related issues (usually in silos), the redundant checks built into it’s systems, lack of visibility into the quality of data as it flows through it’s systems, the significant legal and reputational risk exposure caused by poor quality data, the impact of low quality data on decision making, and the impact of poor quality data on time-to-market. 

Given the data deluge and complexity of the data landscape within firms, there are no quick fixes. However, taking a strategic approach and systematically building the enterprise data quality platform and associated shared service capabilities, will enable firms to overcome the challenges. 

Data Quality Pro: Many of our readers will have a long term strategy for similar senior data quality roles. If we put the technical skills aside for a moment, what are some of the “softer” skills you have had to develop that help a large organisation deliver an enterprise data quality capability?

Jay Zaidi: There are multiple skills that are required to initiate or launch such projects at the enterprise level. The most important ones are influencing skills – to sell the business case and benefits of the program to Senior Management and gain sponsorship for the program, negotiation skills – to negotiate the best deals from vendors and service providers, interpersonal skills – to inculcate a data quality mind set within the enterprise and execution skills – to ensure adoption and timely delivery of the platform capabilities. 

Data Quality Pro: Someone you worked with in the past mentioned your focus on collaboration, how does collaboration feature in your approach to data quality?

Jay Zaidi: Implementing data quality at the enterprise level, is all about collaboration. A small core team of data quality experts in a Data Quality Center of Excellence (DQ-CoE) can not support an entire enterprise by working in a silo. Since the DQ-CoE consults with all LOB’s, it is able to identify patterns and develop data quality solutions for each, which can then be shared across all projects in the enterprise. I am also a big proponent of building repeatable processes and promoting re-use in the enterprise, which results in streamlining operations and reduces time-to-market.

The DQ-CoE team has to partner and collaborate with Data, Application and Technical Architects in each LOB, to assist them in developing sustainable data quality solutions, utilizing expertise it has gained by working across the enterprise. My team has to work closely with the Data Trustees, Data Custodians and Data Stewards to capture data quality requirements and the associated metrics that must be captured. In addition to this, we develop and publish data quality policies and standards on an on-going basis and measure compliance to these across the enterprise. Ofcourse, we work closely with the Enterprise Data Governance team on all such activities, to ensure alignment. 

Data Quality Pro: Let’s shift the discussion to the technical side for a moment. What technical skills do you find yourself drawing on in your enterprise lead role for enterprise data quality?

Jay Zaidi: I have a strong background in database management, data modeling, data analysis and business intelligence, based on an advanced degree in Computer Science and many years of hands-on experience in managing enterprise data. In addition to this, I leverage my Project and Program Management skills and Project Management Professional (PMP) certification, to plan and manage enterprise-level implementations.

Data Quality Pro: Did you plan a career in the data quality profession or has it been more of a gradual, organic journey?

Jay Zaidi: I have always been passionate about data and data management in general. I began my career as a Programmer/Analyst, developing health care solutions and then transitioned into an Oracle database management role. I subsequently moved into Management Consulting and led teams that implemented Commercial-Off-The-Shelf (COTS) applications and designed and developed custom applications for Fortune 500 clients. During the dot com era I led global implementations of Business-to-Business (B2B) and eProcurement applications. So, as you can see, data management has played an important role throughout my academic and professional career. 

Over the past ten years, I’ve had the opportunity to work in all the major lines of business and the horizontal support functions (Enterpries Risk, Audit, Finance, Accounting, etc.) in my current firm and have built strong relationships. I leverage these relationships and my influencing skills to overcome barriers to adoption. 

I did not plan a career in data quality, but was given this opportunity by my current boss, who was aware of my strong data management background and program management skills and confident in my ability to deliver an enterprise-level program. 

Data Quality Pro: Some of our readers will no doubt be planning their own EDQ program or going through those early phases of implementation. What are some of the critical things from your own experience they will need to pay close attention to?

Jay Zaidi: An EDQ Program is a strategic, multi-year initiative that requires on-going investment and organizational support, to sustain it. Your readers need to make this clear to their Program Sponsor(s) and key stakeholders. Implementing such a pervasive program requires a focus on change management, education, training, and technical support. 

Choosing the right data quality COTS product, strategic partners to assist with the roll out and internal resources to operationalize, deploy and manage it, are also critical. 

Data Quality Pro: If I can just jump in and discuss the COTS point. I know you’ve gone through a selection process for data quality products in the past, what tips can you share for making this process as painless as possible?

Jay Zaidi: The COTS product is foundational to the program. I would strongly recommend that your readers do their due diligence prior to initiating this phase of the program. Selecting the right product is an art that requires prior experience in evaluating COTS products and the ability to ask the right questions. The first step in this process is to capture all the data quality use cases within the enterprise. Next, the team should understand the data quality challenges faced by the business, operations and technology teams. This information will be used to develop the evaluation criteria and the logical architecture for the data quality platform. 

For an enterprise level program, you will need a tool set that supports your enterprise’s vision and scales to handle both current and future business needs. The team should conduct industry research and consult with experts before finalizing the tool evaluation criteria, and the RFI and RFP documents. Prioritizing the evaluation criteria and assigning weights accordingly, will facilitate the selection process. I recommend limiting this exercise to the top five to seven leading data quality products. Use a weighted scoring method for the RFI, to select the top three and bring them into your test lab to conduct rigorous functional and performance tests, based on your requirements. 

If you are evaluating products for an EDQ program, some of the areas to focus are the tool’s out-of-the-box data quality capabilities, it’s ability to integrate with other applications in the enterprise, performance and scalability of the product suite, openness of the rules and results repositories, maturity of the product and the vendor’s ability to support your needs. The product road map and total cost of ownership are important criteria as well. 

Make the tool recommendation based on the results of the testing activity, your prioritized evaluation criteria and the weighting model you’ve selected. Consult with industry experts, if you need assistance with this task or require a second pair of eyes to validate your results. I strongly recommend that all key stakeholders be involved throughout the process, to ensure transparency.

Data Quality Pro: In your earlier answer, you mentioned the importance of support. How would you recommend others implement their deployment and support models for an EDQ initiative so that stakeholders can get the maximum benefit?

Jay Zaidi: It is vital that your EDQ team not become a bottleneck in rolling out the program. I would therefore recommend that the data quality platform be deployed using a Federated Architecture. This enables LOB’s to develop data quality solutions in parallel with the enterprise team’s data quality activities. This coupled with a product that has a business user-centric tool and a developer-centric tool, will enable non-technical business users to perform data analysis tasks, hence benefiting all stakeholders. 

This approach reduces the business teams’ dependence on IT and enables them to conduct adhoc data analysis, data quality monitoring and score carding and data quality rules development. Since a significant amount of data quality work is currently done manually or is adhoc in nature, end-users should have the ability to automate their data analysis and monitoring tasks, using the tool’s built-in scheduling capabilities or an external scheduling application. 

The DQ-CoE should provide the necessary on-boarding support, best practices, development framework, training and tool-related documentation to facilitate adoption. 

Data Quality Pro: What are your thoughts on the integration challenges that you can face when deploying an Enterprise tool?

Jay Zaidi: From an Enterprise Data Architecture standpoint, the data quality tool has to integrate with your existing applications, metadata management tool, data integration tool, security tool, business intelligence tool(s) and the Master Data Management (MDM) tool. Typically, this is implemented via adaptors, custom interfaces or web services.

Data quality vendors provide hooks or capabilities to integrate with external systems. However, since each firm manages a portfolio of diverse applications, based on a variety of technologies and architectures, some level of custom development or adaptor configuration will be required, to integrate the tool into the enterprise. This should be investigated during the tool evaluation phase, to determine which product(s) can best support your enterprise’s systems integration requirements. I would like to note that data quality plays a critical role in MDM implementations, so integration between the EDQ and MDM tool must be tested thoroughly, if MDM is a priority or there are plans to implement MDM in future.

Something else that should be high on your list is an ability to access EDQ-related metadata, EDQ results, EDQ metrics, transaction and exception logs and other relevant data, stored in the tool’s data repositories. Many COTS products have this kind of information trapped within a proprietary database, so be sure you can get at it easily. It will become vital to your ongoing EDQ efforts.

Data Quality Pro: What do you recommend for getting the business bought into the program as they are obviously critical to long-term success?

Jay Zaidi: Data is pervasive and continues to grow rapidly in every business. Every LOB and business organization deals with data quality challenges. Based on my interaction with other professionals in the industry, it is clear that many companies still utilize highly manual and adhoc solutions to address data quality issues. I also see a lack of data quality frameworks that can be utilized to standardize the definition of data quality requirements and the associated metrics, to enable consistent measurements and an apples-to-apples comparison of data quality across the enterprise. 

We assume that business users are as savvy and educated about the various tools and technologies as we are. But this is not the case. They are subject matter experts in their specific business domains, but often lack awareness about the latest developments in data management. Educating users on the value of data quality to their organization and training them in the tool set are important considerations. My experience has shown that a lack of education and tool training can often result in user frustration, which if left unchecked can have a negative impact on the program and ultimately result in it’s failure. Constantly monitoring feedback from end-users and proactively addressing issues is important. 

Risk Management, Governance, Compliance, Audit and Finance organizations have a vested interest in supporting and adopting the enterprise data quality program. To gain traction and get some quick wins, I suggest targeting these organizations early in the program. 

The ability to automate data quality monitoring and score carding activities, share and re-use data quality rules across the LOB, deploy generic data quality services, and proactively manage data quality typically resonate with the user community, especially the business teams. 

The EDQ platform should be bundled with consulting, technical support and best practices documentation and the DQ-CoE should provide documentation templates, design patterns and development frameworks, to reduce time-to-market. To reduce barriers to adoption, I’d recommend that the EDQ platform, bundled services and best practices be provided to the LOB’s free of charge. In today’s environment of shrinking budgets this makes a highly attractive proposition.

Data Quality Pro: I particularly like your suggestion of creating re-usable components and rules that can be shared amongst the different LOB’s to accelerate take-up. What type of rules and components would you recommend from experience?

Jay Zaidi: Most of the leading data quality tools provide the ability to create re-usable components and rules that can be shared across different LOB’s. This requires some pre-work on the part of the designer, to understand the use cases and ensure that the components are designed in a generic manner and can be leveraged by multiple applications. I would recommend that you create re-usable rule sets, web services, data validation, data standardization, data cleansing and normalization components, based on generic data quality requirements. For example, in our industry, Counter party data is expected to be received in a certain format and adhere to a set of validation checks. By creating generic web services to vend to various applications, you can deliver a considerable amount of data quality management capability, without having to build or change a multitude of different applications and interfaces.

Data Quality Pro: Are there any specific change management techniques or methodologies that you feel our readers will find useful in helping their organizations to improve their awareness and attitude towards data quality?

Jay Zaidi: Human beings are resistant to change, so implementing any program that requires significant change in behavior or requires learning new tools and techniques, will generate resistance. For an enterprise level program, you will need to cater to the needs of technical and non-technical clients, who have a limited knowledge of data quality management and best practices. 

Be sensitive to the fact that non-technical users adapt well to tools that are intuitive and easy to use, so factor this into your product evaluation criteria. In addition, you should focus on educating end-users on the value an enterprise data quality program brings to their organization. In particular, I would recommend focusing your message on automation, process efficiency, reduction in operational incidents, transparency into data quality issues and the user’s ability to proactively manage issues using the data quality platform. 

Data Quality Pro: What is the current appetite for implementing ”Holistic Data Quality” – a term you coined? 

Jay Zaidi: If you take a step back and analyze the Information Supply Chain (ISC) of any organization, you will notice that business is conducted left to right, in that ISC and data tends to flow that way, but we process and manage it in horizontal silos. I’ve always had a problem with this approach – since what this does is ends up impeding the flow of business, introduces data quality issues during hand-offs between the silos, impacts time-to-value, doesn’t provide a “holistic” view of the data and it’s quality and is very resource intensive, due to the amount of redundant activity that is conducted in these silos. “Holistic Data Quality” is a term that I coined to highlight the fact that data quality should not be evaluated or managed in silos, but requires a holistic cross-silo approach. I would suggest that data management practitioners take a step back and evaluate their current approaches to data quality management, to determine if a holistic approach provides a better alternative.

Data quality is gaining attention lately, due to some high profile incidents that resulted from poor data quality. As I have mentioned earlier, the focus is now on addressing systemic issues. New legislation and oversight bodies are being instituted to monitor and address systemic issues related to data quality, particularly with multi-nationals and government regulated firms. Large firms in the Financial Services sector have noticed this and are taking steps to implement “holistic data quality”. I’d like to conduct a separate interview with you on the impact of new regulations such as Dodd-Frank on a firm’s data management practices and why they should be seriously consider taking a “Holistic Data Quality” approach, in addition to other strategies.

Data Quality Pro: How can readers make data quality relevant to business heads and other senior executives who may be, quite rightly, far more interested in how the business is performing as opposed to how the data is performing?

Jay Zaidi: In order to show Executives and LOB heads how they can utilize data quality metrics to identify data anomalies and take proactive steps to address them, you need to conceptualize and develop an Enterprise Data Quality Dashboard. Initially, your prototype may be based on synthetic data that simulates real life metrics which are meaningful to senior executives. Your dashboards must then depict the trending for data quality metrics across several dimensions, perhaps over a three month period. A side-by-side comparison of the quality of various data elements at different times can then be provided. 

Most dashboard tools would allow you to include data quality thresholds and for this a traffic light concept (Red, Yellow, and Green) can be very useful to easily indicate the health of the data. The dashboard should also provide drill down capabilities to enable users to quickly visualize any outliers. 

From personal experience, providing Executives and LOB heads a visual representation of the data quality metrics and the ability to identify anomalies and patterns, can significantly help to create excitement and sponsorship within the senior management ranks. 

In the past, we have utilized the feedback received on our prototype, to design the production scorecards and dashboards. As with any project of this nature, it is important to remember that Executives and LOB heads require summary level information, with the ability to focus on outliers, and quickly get to the root cause. 

Data Quality Pro: Finally, what lessons can you share with others who are about to embark on an Enterprise Data Quality initiative?

Jay Zaidi: Implementing any enterprise level program is a challenging assignment and hence one must embark on this journey with a clear vision, sponsorship at the highest level of the company, a commitment for long term financial and political support from senior management and an appetite for change. 

Given the pervasive nature of data, implementing data quality across an enterprise requires multi-tasking, strategic planning, promoting end-user adoption via education and training, attention to detail, and influencing skills, to motivate the end users and support teams. Active risk management is critical to the success of the project. 

After making the tool selection, designing your Solutions Architecture and finalizing your roll out strategy, attention will quickly shift to program execution. This is a critical phase and requires strong leadership, attention to detail, a robust risk management regime, meticulous planning and a team of seasoned resources that have prior experience in deploying COTS products at the enterprise level. Given the highly specialized nature of a data quality solution, the COTS product vendor and it’s professional services staff will play a critical role during this phase, so getting them on-site and working within the larger execution team during all phases, is essential. Ensuring that the vendor has “skin-in-the-game”, with the same level of commitment and sense of urgency to the program as you do, reduces risk. 

Deploying narrowly scoped data quality solutions (low hanging fruit) early in the program’s life will ensure buy-in, end user adoption and momentum for the program. Focus should be on adding business value quickly, so the DQ-CoE should target high business value projects, that can benefit from EDQ solutions. It has been my experience that time-to-market is extremely important to the business. I suspect this applies to most companies, so implementation teams must focus on this. Pre-built templates, design patterns, repeatable processes, re-usable components, re-usable rules, and an iterative approach should be utilized to reduce time-to-market. Key stakeholders will lose interest in the program, if it doesn’t add business value quickly. 

As I have stated before, implementing a data quality program at the enterprise level should be treated as a marathon and not a sprint. I would encourage firms to consider developing such a program, if they haven’t already and to systematically build it over time. As you can see, I am very passionate about this and am very interested in hearing from your audience, and in helping them address their data quality challenges.

Data Quality Pro: Thanks for sharing your time Jay, best of luck for 2012.

Jay Zaidi: It’s my pleasure Dylan.

Jay Zaidi - Enterprise Information Management Leader

Jay Zaidi – Enterprise Information Management Leader

Jay Zaidi

Jay Zaidi is an astute, hands-on, versatile and results-oriented leader with proven success in Enterprise Data Management, Strategic Planning, and Program Management. 

He is passionate about solving multidimensional problems. During his professional career, Jay has conceptualized and led business transformation and change management programs in the Financial Services business vertical. 

He has led global data management projects to address Regulatory Compliance, Risk Management and Operational challenges. He consults with and influences all levels of management and works to bridge gaps, facilitate communication and develop integrated business solutions. Proven success in strategic guidance to leaders in Fortune 100 firms. 

Jay can be reached via his LinkedIn Profile.