As we discussed in the previous article in this series, there seems to be an inherent contradiction in executive suites these days between the perceived importance of data quality and the actual
commitment of resources to data quality initiatives. While executives increasingly endorse data quality as a valuable business asset, data quality software is often evaluated in terms of product
features and functions rather than as a business solution that provides strategic advantage. As a result, organizations are not realizing the substantial benefits that can be gained from their data
quality investment.
One of the major reasons for this trend toward data quality software being commoditized is the selection and implementation process.
Over the past few years, we have seen a dramatic increase in the time and expense required to evaluate and select data quality software. It has become a drawn-out, expensive process, often
involving extensive questionnaires listing hundreds of software features and functions. Completing and reviewing these questionnaires can be a tedious, time-consuming task – for the client as well
as the vendors. And despite all the time, effort and expense invested in making the “best” selection, the software may still end up on the shelf soon after the initial conversion.
What is the reason for this non-productive use of money and resources? It has been my experience that in many cases, the selection and implementation process is managed by someone who is not truly
familiar with the unique demands and intricacies of data quality processing. At best, they may have performed only two or three previous data quality implementations. Often they do not have even
this limited experience. The result is:
- The evaluation process does not focus attention on identifying the client’s business-driven data quality needs, and how each vendor’s offering relates to these needs
- The standard list of features and functions included in the qualifying questionnaire may have little to do with how the client will actually use the data quality software selected
- During the proof of concept, the competing vendors’ solutions are often ‘tested’ by using a sample of the client’s data that is not valid in size or composition to produce reliable and
meaningful results - The data quality software is implemented on a project-level basis and typically does not take full advantage of the robustness of the selected solution’s enterprise capabilities
When these events occur, the all too familiar cycle of data quality software under-performance is repeated. Although the client has made a tremendous investment of time and money to evaluate and
choose the most suitable solution from the competing vendors’ offerings, there remains significant risk that the data quality solution selected will not generate the productivity and ROI
projected. Because the evaluation criteria and process were flawed, the client becomes dissatisfied and the data quality software is relegated to the shelf.
An Alternative: A More Focused Approach
Let’s consider the effect on the overall process of adding an independent data quality specialist to the integration selection and implementation team. By “specialist,” I mean someone who has a
minimum of five years’ data quality processing experience, is familiar with the different data quality offerings on the market and has performed at least 20-30 data quality implementations.
An independent data quality specialist brings important insight and guidance to the evaluation and selection process, and becomes a critical bridge for ensuring that both IT and business
requirements can be achieved with the data quality solution ultimately selected.
Following are four important ways that a specialist can strengthen the software selection and implementation process:
1. Focusing the Process on the Client’s Needs
An independent data quality specialist understands how data quality software works and how it applies to solving business problems. With this expertise, the specialist can help focus the evaluation
and selection process on the client’s own situation, in terms of:
- How the data quality solution will be used
- Appropriate accuracy levels to be achieved to meet the client’s real business needs
- IT resources required for a robust implementation and ongoing maintenance
- Realistic implementation schedules in relation to the client’s own time constraints
Focusing the evaluation effort on the particular client’s unique, business-driven data quality objectives is much more efficient and meaningful than reviewing an omnibus list of features and
functions. It will help ensure that the solution selected truly meets the client’s needs. It will also streamline the process, reducing the cost and time-frames required.
2. Generating Creative, Client-Focused Problem Solving
To take the client-focused approach a step further, the data quality specialist can position part of the RFP as a challenge based on the client’s unique data quality needs and objectives (the
bulleted items listed under #1). Each vendor will be asked to present their best solution for maximizing the client’s immediate and ongoing data quality performance, within the given parameters.
For the client, this is certainly a more productive strategy than comparing checked boxes on competing vendors’ questionnaires. It will also provide a very effective way of differentiating the
vendors in terms of their software’s capabilities and their problem-solving skills.
3. Selecting an Appropriate Sample for Testing
Proof of concept is a critical phase of most software evaluations. But a proof of concept is only as valid as the test data used. A data quality specialist can help ensure the validity of a proof
of concept by selecting an appropriate test sample.
Samples provided by clients are often much too small to be statistically significant. The results of tests performed with such samples cannot be generalized to the implementation as a whole and are
therefore relatively meaningless.
Potentially even more damaging, a small data sample can be rigged by a vendor to make their data quality solution look better. The solution demonstrators can tune their software to the specific
parameters of the small sample provided, skewing the results. The only way to perform a realistic test of that software is to run it on another sample of data without any additional tuning.
4. Providing a Robust Implementation
Too often during major systems integrations, the data quality implementation is considered only from the standpoint of the integration project itself – migrating the data from Point A to Point B.
Functions such as data auditing and ongoing quality maintenance may be regarded as extras and not really critical to the integration process. The project leader may therefore omit the installation
of these important data quality functions in the interest of saving time and resources.
A data quality specialist will recognize the full functionality and capabilities of the software selected, and approach the implementation from the standpoint of ensuring ongoing data quality.
During the evaluation stage, the data quality specialist identified how the software would be used, and determined the appropriate quality levels required to meet the client’s business needs.
During the implementation, they can help customize the installation to the nature and data quality demands of the client’s business.
The data quality specialist can help to ensure that:
- The implementation is robust and takes full advantage of the software’s functionality and capabilities in relation to the client’s needs
- Appropriate data quality accuracy levels are established, based on the data’s business uses
- A program for regular data quality audits and ongoing maintenance is established
When these factors are all built into the initial implementation, the data quality solution is likely to meet the client’s expectations and quickly begin to generate ROI. This data quality
software won’t end up on the shelf.
5 Key Questions for Selecting a Data Quality Vendor
This article is essentially about turning the perceived business value of data quality software into reality. There are a number of highly qualified data quality vendors in the marketplace today.
Many of them can deliver a competent solution that – if properly implemented and managed – can ensure the ongoing quality and integrity of your organization’s data asset.
However, the ultimate problem is still: With all the data quality vendors in the market today, how do you identify the top performers? There are some key questions to ask that will help determine
their level of expertise, confidence in their software and commitment to customer service. Following are five of the critical questions, with a brief explanation of the kind of response you should
expect:
-
What is the average tenure of your delivery consultants? (A Senior Delivery Consultant should have at least five years’ experience and have participated in at least 20 data
quality implementation projects) -
Can I speak with a customer who licensed your software at least three years ago and is still working with you? (This will show you how the company treats its clients after the
initial licensing income dwindles to a lower, maintenance level) - Can you provide an overview of your data integration methodology? (The methodology should include:
-
- Determining how the data quality suite will actually be used
- Defining the accuracy levels that will be needed to meet the client’s business needs
- Profiling the data to determine existing quality levels and identify potential problem areas that must be addressed
- Performing a test run using an appropriate statistically significant data sample prior to the full integration)
-
Other than data quality integration, what business lines does your company focus on? (Remember, you’re looking for an expert in data quality. If their business is diversified,
they may not have the focus to provide the expertise you need.) -
How long has your company been in business and how many data quality implementation projects have you performed? (Clearly, the more experience, the greater the credibility. One
hundred projects is a minimum figure. Five hundred is establishing integrity. One thousand or more, and you probably have a true data quality specialist.)
Fulfilling the Data Quality Promise of Your Next Systems Integration
Although the business value of data quality is increasingly being recognized in executive suites and boardrooms, broad commitment to enterprise data quality initiatives in terms of budget and
resources is still lagging. So when an organization is successful in securing the necessary resources to select, license and implement a data quality solution, it is especially disappointing if
that data quality software ultimately becomes shelfware.
The shift from software to shelfware can be prevented. One way to break the shelfware cycle is to have a data quality specialist participate in the software selection and implementation process.
The specialist’s expertise can focus the selection process on the client’s own data quality needs and objectives, enhance the validity of proofs of concept, provide grounds for differentiating
the vendors, and ensure a robust, customized implementation.
I encourage you to keep these factors in mind when planning your next data integration or data quality initiative. If your team’s initial plan does not include a data quality specialist on the
project team … insist on it. It will pay off in the performance and ROI of your data quality solution.
What’s Next …
Now that we’ve set the necessary foundation for effective customer data integration by outlining the components of a closed-loop data management environment, discussing the importance of corporate
commitment to enterprise data quality and presenting an alternative approach for selecting a data quality solution, in our next article we will address how to effectively remove risk from
mission-critical data conversion projects.
Whether driven by mergers and acquisitions or implementations of enterprise application systems, customer data integration projects present complex challenges to even the most technology-savvy
organizations. Recognizing that customer data is an essential component of most business-driven conversion projects, organizations should establish an incremental conversion strategy that is
uniquely designed to ensure the reliability of customer data throughout the conversion process. The next article in this series will present a proven methodology for successful data conversions
regardless of time, cost and resource constraints.
