What Is the Number One Problem in CDI?
If you had asked me that question a couple of years ago when CDI was a fledgling technology, I would have guessed that the toughest problem was based in technology, namely the vast quantities of data stored in mainframe systems and legacy applications that do not have open APIs for accessing the data. And I would have been dead wrong.
Turns out that the vast majority of companies have already found a need to work with that data in some fashion, and have crafted solutions for accessing the systems, from direct DB2 database reads to custom connectors and query systems. At the same time, a number of vendors have carved out profitable niches with products that establish direct connections between clients and the host server, products that channel connections through the Web server, and even real time bi-directional read-write connectivity to mainframe systems.
Today, based on working with clients on dozens of CDI projects from business justification through implementation, I would say that the number one problem is in the area of executive sponsorship. This has three practical aspects, from the authorization, scoping and funding of the project (sponsorship), to the issue of data “ownership” and project responsibilities (governance), through the data stewardship function and the ongoing CDI efforts which must be maintained for the long-term success of the initiative (project management).
There is today a healthy tension between the need to upgrade existing systems and implement superior technologies such as web services and a true SOA on the one hand, and the need to cost justify every project on the basis of a rigorous ROI analysis. The problem for CDI is that it is a true enterprise application. There is always a “community good” aspect to the project, meaning that some business units will bear the majority of the cost and responsibilities while others will get a “free ride”, enjoying the benefits of accurate and complete customer data without incurring much of the pain or cost.
To move the project off the white board and into development, true executive sponsorship is called for. We all know that sometimes “executive sponsorship” is simply where the buck stops when disputes must be resolved. For a CDI project to succeed, indeed for it ever to really get started, true leadership is required. This means senior management is engaged, equipped to express a clear vision of the business drivers behind CDI, and actively and vocally insisting on alignment between the business owners in the principal divisions and departments affected.
Sponsorship
CDI is an enterprise-wide initiative. In planning for a CDI project, it is absolutely necessary to plan the end-state including data schema, hardware configuration, integration architecture, messaging and middleware environments. Only then can short-term tradeoffs be evaluated in context. Only then can the business owners be assured that work done on smaller, short-term projects is in alignment with, and reusable with, the larger enterprise CDI. Putting an organizational framework around the project is required in order to even discuss how to begin, and this is often not something that already exists.
Governance
Governance is about internal coordination, and is required to address enterprise issues such as creating internal alignment, who “owns” the data, how new enterprise functions such as data stewardship and CDI center of excellence are staffed and funded, ensuring the appropriate participation and compliance among the various business units, providing for objective measurement of deliverables, creating the project budget, managing partners, staffing the project, creating the enterprise business case, ROI, TCO and securing project funding.
Project Management
Project management focuses on coordination of tactical deliverables between various teams focused on the target architecture, initial data load, hardware scalability and performance, business process management and automation, manual workflows and UI design, data quality and data enhancement, integration architecture, system analytics, PSR, operations, configuration and testing, retail and commercial CDI combination or interaction, and documentation. Because the scope of a CDI project is typically very broad, involving many systems and many business units, project ownership and responsibility need to assigned up front. Authority and accountability, with both communicated to all the affected parties.
Vendors and SIs can help with project management, but the sponsorship and governance need to be driven by the company implementing the CDI solution. Companies can and should rely on their vendors and consultants to add value with economic analysis to feed ROI models and TCO analysis, best practices from existing CDI centers of excellence, initial staffing estimates for various aspects of the project, and many other elements. Companies considering CDI projects need to be prepared to provide clarity around the governance structure for the project, and answers to basic questions such as who is developing the business case, what is the funding process, what are target timelines for specific deliverables required to cost-justify the project, what internal staffing resources can be made available and when, and who owns requirements-gathering for the various aspects of the project.
CDI projects typically encompass a variety of smaller projects, each with its own audience and interest groups. In addition, resolving enterprise-wide project ownership issues is complex in its own right, and involves accommodating a wide variety of needs and requirements. Communication is a critical part of the solution, from the clear leadership message on the CDI initiative emanating from a CEO or COO, to requirements discovery among all affected business units, to setting expectations correctly among the end-user community, to defining – and publishing – and ensuring compliance with – the newly minted processes for working with customer data.
Turns out that the vast majority of companies have already found a need to work with that data in some fashion, and have crafted solutions for accessing the systems, from direct DB2 database reads to custom connectors and query systems. At the same time, a number of vendors have carved out profitable niches with products that establish direct connections between clients and the host server, products that channel connections through the Web server, and even real time bi-directional read-write connectivity to mainframe systems.
Today, based on working with clients on dozens of CDI projects from business justification through implementation, I would say that the number one problem is in the area of executive sponsorship. This has three practical aspects, from the authorization, scoping and funding of the project (sponsorship), to the issue of data “ownership” and project responsibilities (governance), through the data stewardship function and the ongoing CDI efforts which must be maintained for the long-term success of the initiative (project management).
There is today a healthy tension between the need to upgrade existing systems and implement superior technologies such as web services and a true SOA on the one hand, and the need to cost justify every project on the basis of a rigorous ROI analysis. The problem for CDI is that it is a true enterprise application. There is always a “community good” aspect to the project, meaning that some business units will bear the majority of the cost and responsibilities while others will get a “free ride”, enjoying the benefits of accurate and complete customer data without incurring much of the pain or cost.
To move the project off the white board and into development, true executive sponsorship is called for. We all know that sometimes “executive sponsorship” is simply where the buck stops when disputes must be resolved. For a CDI project to succeed, indeed for it ever to really get started, true leadership is required. This means senior management is engaged, equipped to express a clear vision of the business drivers behind CDI, and actively and vocally insisting on alignment between the business owners in the principal divisions and departments affected.
Sponsorship
CDI is an enterprise-wide initiative. In planning for a CDI project, it is absolutely necessary to plan the end-state including data schema, hardware configuration, integration architecture, messaging and middleware environments. Only then can short-term tradeoffs be evaluated in context. Only then can the business owners be assured that work done on smaller, short-term projects is in alignment with, and reusable with, the larger enterprise CDI. Putting an organizational framework around the project is required in order to even discuss how to begin, and this is often not something that already exists.
Governance
Governance is about internal coordination, and is required to address enterprise issues such as creating internal alignment, who “owns” the data, how new enterprise functions such as data stewardship and CDI center of excellence are staffed and funded, ensuring the appropriate participation and compliance among the various business units, providing for objective measurement of deliverables, creating the project budget, managing partners, staffing the project, creating the enterprise business case, ROI, TCO and securing project funding.
Project Management
Project management focuses on coordination of tactical deliverables between various teams focused on the target architecture, initial data load, hardware scalability and performance, business process management and automation, manual workflows and UI design, data quality and data enhancement, integration architecture, system analytics, PSR, operations, configuration and testing, retail and commercial CDI combination or interaction, and documentation. Because the scope of a CDI project is typically very broad, involving many systems and many business units, project ownership and responsibility need to assigned up front. Authority and accountability, with both communicated to all the affected parties.
Vendors and SIs can help with project management, but the sponsorship and governance need to be driven by the company implementing the CDI solution. Companies can and should rely on their vendors and consultants to add value with economic analysis to feed ROI models and TCO analysis, best practices from existing CDI centers of excellence, initial staffing estimates for various aspects of the project, and many other elements. Companies considering CDI projects need to be prepared to provide clarity around the governance structure for the project, and answers to basic questions such as who is developing the business case, what is the funding process, what are target timelines for specific deliverables required to cost-justify the project, what internal staffing resources can be made available and when, and who owns requirements-gathering for the various aspects of the project.
CDI projects typically encompass a variety of smaller projects, each with its own audience and interest groups. In addition, resolving enterprise-wide project ownership issues is complex in its own right, and involves accommodating a wide variety of needs and requirements. Communication is a critical part of the solution, from the clear leadership message on the CDI initiative emanating from a CEO or COO, to requirements discovery among all affected business units, to setting expectations correctly among the end-user community, to defining – and publishing – and ensuring compliance with – the newly minted processes for working with customer data.
1 Comments:
Funding CDI is the biggest problem at my bank, but reconciling all the different departmental standards for data models and metadata is also a major headache. The wonderful thing about standards is that there are so many of them to choose from ;-)
Post a Comment
<< Home