Pick Your Partners:
Evaluation Criteria for CDI-MDM Systems Integrators
Published: April 1, 2007
This article focuses on analysis of product strategies and best practices rather than the vendor press releases.
Published in TDAN.com April 2007
The intent is to provide ongoing advice for enterprise data architects, data stewards, CIOs, CTOs and CDI-MDM project leads in helping develop their CDI-MDM strategies – via independent, authoritative, and relevant analysis.
Preparing for CDI-MDM talent shortages now is the savvy way to avoid the systems integrator "money pit" later. As noted in a recent CDI-MDM newsletter article, Is Now the Time for 'Regime Change' in Your Incumbent Systems Integrator(s)?, historically there has been economic and political pressure to stay the course with your incumbent systems integrator (SI) for enterprise-scale projects. However, what if your preferred SI partner does not have any project experience in the rapidly evolving CDI-MDM technology arena or is tremendously short staffed in this area? What do you do? In this article, we provide a synopsis of how one might evaluate the capabilities of your "new" potential SI partners.
Bottom line: Given the strategic nature of CDI-MDM projects, the capabilities of partner SIs must be given close scrutiny - not only in an effort to contain costs, but also to insure the success of this vital infrastructure investment. As stated in our September 2005 DM Review column, "Avoiding the CDI 'Money Pit' - Pick Your (Systems Integration) Partner Wisely," preparing for CDI-MDM talent shortages now is the savvy way to avoid the CDI-MDM "money pit" later.
"Top 5" Technical Evaluation Criteria for CDI-MDM Systems Integrators
The focus of this article is to outline the "top 5" technical evaluation criteria used by large IT enterprises to identify their SI partner(s) to help implement their next generation of customer data integration (CDI), master data management (MDM), and/or data governance initiatives.
During first quarter of 2007, the CDI-MDM Institute is wrapping up the final stage of a large scale survey of 350+ enterprises and their views on how to identify, evaluate, and manage SI partners for CDI-MDM initiatives. To sample this 10-minute survey ("take a survey, get a scorecard"), follow this link.
It is assumed that most SIs have the same underlying business model whereby "partners" maintain strengths in key areas of expertise (industry-specific or technology-specific) and maintain varying degrees of competency in others. In any case, the CDI-MDM practice partner is able to draw upon resources from centers of excellence (CoEs) or centers of competency (CoCs) maintained by other partners. As expected, the CDI-MDM partners often evolved from the business intelligence, ERP and CRM specialty practices. And just as often, the multi-disciplinary teams assembled to work on your project(s) will be pulled from these same areas: enterprise architecture planning, especially SOA experts; enterprise data modeling; legacy application reengineering; data conversion and application migration; data quality; EAI middleware and BPM experts; analytics and ETL experts; testing and QA; systems infrastructure management; performance engineering; and ROI and cost-model generation (business value articulation).
Top Five Technical Evaluation Criteria
What are the "top 5" technical evaluation criteria for the services of systems integrators on CDI-MDM projects? An overarching consideration is that we need to be flexible in our understanding that few SIs have experience with one major CDI-MDM project - let alone more than one in our industry. "In-flight" projects are no substitute for experience, but do represent a vote of confidence by your industry in that SI's capabilities as well as a commitment by your SI partner that they are committed to your industry.
According to the CDI-MDM Institute, the top five evaluation criteria are:
Extensible Data Governance Methodology & Accelerators
Too many times, our clients report to us that "our SI doesn't understand us," which translates to "we asked for a CDI or MDM proposal and what our SI gave us was a rehash of their enterprise data warehouse method." Yes, enterprise data warehouses (EDWs) and business intelligence systems in general are often feeder systems into a CDI-MDM solution; however, CDI-MDM solutions are much more near real-time than the batch-style data warehouses that are endemic in the corporate IT world. Moreover, CDI-MDM solutions both straddle and bridge the front- and back-office systems to provide shared services for master data CRUD functions such as enterprise-wide replicated customer address/marital/phone changes. A key area of differentiation for methodologies is the level of integration between the visually oriented process mapping tools and the actual "process generators" that feed the rules engine to drive the master data policies captured in the methodology. Increasingly, we can expect the best-of-breed CDI-MDM software vendors to promote this notion of "data governance console/methodology" to use their tools as a front-end design tool to the much more cumbersome "Mack truck" approach of IBM WCC and Oracle-Siebel UCM. Given the current lack of integration among the various CDI-MDM design and methodology tools with the actual CDI-MDM platforms, evaluators of SI's offerings need to pay close attention to the level of integration today as well as the future for any such data governance tools on the SI's road map (or how the SI plans to leverage future tools from the software partners). See also our recent white paper on data governance.
Industry-Specific Data Model Experience & ETL Mappings
Data models are often the top technical evaluation criteria of CDI-MDM software solutions, so it follows that the technical skills and expertise of the SI candidates must match your data model center of gravity - e.g., CSC Hogan, IBM Banking Data Warehouse, Oracle Trading Community Architecture, Siebel CRM, etc.
Moreover, the SI must have the prerequisite understanding of the data models of the key feeder systems into the CDI-MDM hub. That is, the SI must be able to map the primary data sources into the hub either through leverage of ETL tools with their associated mappings, or by bringing to the engagement their own custom mappings of Oracle, PeopleSoft, SAP, Teradata, etc. application data models.
SOA Architecture Experience & Accelerators
As noted in the extensible methodology requirement, the architecture design point for third generation CDI-MDM solutions is not batch and is not isolated in either the operational or analytical (left brain/right brain) division of our enterprise. Rather, the design point of such CDI-MDM software is typically shared services/policies/processes - a.k.a. service-oriented architecture (SOA) and the component-based application approach. Quite commonly, a commercial CDI-MDM product is often an IT organization's first "foray into SOA." And like anything major attempted for the first time, it mandates an SI partner who has experience. Most SIs take an approach wherein they: (a) provide their own SOA framework as an accelerator, and (b) leverage the SOA frameworks provided by the mega vendors (IBM WBI, Oracle-Siebel UAN, SAP NetWeaver) as well as the best-of-breed CDI-MDM platforms (Initiate Systems, Purisma, Siperian, et al). In any case, the SI has to keep up to date regarding the current and future capabilities of the CDI-MDM software partners in terms of their support for process flow integration, which is why the notion of "process hubs" as promoted by IBM and SAP is increasingly relevant - not just to "Type A" early adopters. Experience with the underlying rules engine of commercial CDI-MDM software and/or BPEL is also required. See Dysfunctional Data Hubs: 2006-07 Strategic Planning Assumptions for CDI-MDM & Business Services.
CDI-MDM Product Experience
As IT executives may have noticed, costs for CDI-MDM product-specific consulting are often out of line with other rank-and-file IT skill sets. During 2007-08, acute skill shortages in mega vendor products such as IBM WCC, Oracle MDM (a.k.a. Siebel UCM), and SAP NetWeaver MDM will significantly exacerbate project costs. While "vendor product certification" is good, "product experience" is also vital. Notably, such key features as hierarchy management support vary widely from one CD-MDM product to another. That key component, along with experience with Acxiom and D&B, etc. hierarchy management solutions. Ideally, the SI partner brings process templates and tools (e.g., product evaluation matrix, gap analysis, etc.) to assist in the evaluation and selection of "build vs. buy" as well as "CDI-MDM product selection" (if "buy").
CDI-MDM Project Experience
Additionally, shortages of CDI-MDM project and process skill sets will drive up costs for enterprise data architects, data stewards and other individuals with strong affinity for data governance. Clearly, demand will outstrip the market supply for individuals with actual experience as few individuals have one project under their belt, let alone more than one project. This will create demand for such SI innovations as "rent-an-architect" and "rent-a-data-steward." The SI partner should be expected to assist in the project ROI justification as a standard matter via pre-loaded economic models to help IT and business leadership engage in CDI-MDM justification processes. For more discussion, see CDI Increasingly Seen As 'Career Durability Insurance.'"
Key Business Factors in SI Evaluations
The article Is Now the Time for 'Regime Change' in Your Incumbent Systems Integrator(s)? highlighted why SIs are vital to the success of CDI-MDM projects:
In any discussion of SIs, however, there still remain three key business factors that need appropriate weighting in any evaluation process:
The Importance of SIs to CDI-MDM Success
To rhapsodize on the current "group think" of the CDI-MDM Advisory Council regarding the importance of evaluating new SI partners for their CDI-MDM projects, one could summarize their advice as:
Alternatives abound to your historical SI partners. During 2Q2007, the CDI-MDM Institute will release its MarketPulseTM report titled "CDI-MDM Systems Integration Services: 2007-08 Market Review & Forecast," which will include ratings for both first- and second-tier consultancies as well as evaluation criteria and other findings based on surveys of the CDI-MDM Institute Advisory Council. For a preview of the SIs and consultancies to be reviewed, browse the Top 50.
Bottom Line (Redux)
Given the strategic nature of CDI-MDM projects, the capabilities of partner SIs must be given close scrutiny - not only in an effort to contain costs, but also to insure the success of this vital infrastructure investment. Preparing for CDI-MDM talent shortages now is the savvy way to avoid the CDI-MDM "money pit" later.
Recent articles by Aaron Zornes
Aaron Zornes - Aaron Zornes is the founder and chief research officer for the MDM Institute. He is also conference chairman for the “MDM & Data Governance Summit” conference series, which is the most widely attended professional conference focused exclusively on MDM and Data Governance (London, New York City, San Francisco, Shanghai, Singapore, Sydney, Tokyo and Toronto).