When this is looked upon purely from a software perspective there's not much difference. Sure, there are cost savings associated with the reduction in the storage requirements. There might even be some increased performance in dependent applications due to the reduced volume. However this is hardly a justification for the investment that a typical data quality initiative requires. This is particularly inconvenient considering most of the investment is in software and other technology related resources.
However consider the impact of a data quality project which consolidates customer data from a business perspective and see a different side of things. Consider the benefits of less, unnecessary, possibly inaccurate customer data.
- fewer mailings to reach the same customer providing a direct cost savings
- fewer mailings to reach the same household providing a direct cost savings
- fewer mailings required overall providing a direct cost savings
- fewer failed mailing attempts due to address validation providing a direct cost savings
- fewer customer service requirements due to single view of the customer providing a direct cost savings
- more accurate perspective of customer product portfolio providing a direct increase in marketing opportunities
Now (re)consider the substantial impact that can be realized from a consolidation effort. Furthermore as long as data quality initiatives are implemented into ongoing operational data services, these cost reductions extend into the future producing benefits in the long term. This further justifies the cost of implementing data quality services into an organization as a long term solution.
This is why it is critical to the success of a data quality project to have clear goals that are aligned with a business initiative.
However this is not the end of the line when it comes to ensuring success. To do this you have to start with a goal like the ones listed above and define ways in which these types of goals can be measured.
For example the first bullet point is a data quality goal tied to the business initiative of reducing duplicate customer data. To support this a data quality matching process can be defined that uses criteria to identify redundant customer transactions and consolidate them into a survivor record. The affect the data quality initiative has on this business process can be measured in terms of the reduction in total mailings required to complete a marketing campaign. More importantly, it can be measured in terms of a reduction in total dollars required to fund the new and more concise direct mailing campaign. Now the data quality process and its results can be linked directly to a reduction in budget. Clearly metrics like these make it obvious that a data quality initiative that merely reduces data has a tremendous amount of value.
If you define a list like this with business stakeholders driving the process, before the data quality project is implemented, there will be a clear path to success as well as an easy way to quantify it once the solution is deployed!
Nice post William, in these climates I think it's impractical to embark on any kind of DQ activity (other than compliance) without a clear ROI model in place.
ReplyDeleteThe key as you say is to have a firm understanding of what you expect to achieve, this can be done using a sample or lightweight trial beforehand but you're absolutely right, when the project kicks off, sponsors do expect to see results.
Can I also welcome your blog to the global DQ community, I'll update our records and add you to the next Data Quality Pro Blog roundup.
-Best of luck with the site,
Dylan
Excellent post!
ReplyDeleteI have seen too many data quality projects try to get by with a weak case such as “fix the bad data” or “reduce the number of erroneous records” without linking to a tangible business case or clear definition of ROI.
When a data quality project is performed this way, it will always been viewed as either a failure or at best a non-event that no one even notices no matter how well the data quality was actually improved.
Therefore, I completely agree that you have to present a solid business case that illustrates the negative impact of poor data quality on decision-critical enterprise information, such as incorrect business decisions, bad customer experiences and ultimately lost revenue.
I was in a similar discussion this afternoon! There's a subtlety to your post that is important, and I think Dylan's comment might miss it a little bit.
ReplyDeleteThere are two extremes on the "customer data quality value proposition" front: 1) quantify direct business benefits that the project will provide and forecast (and later measure) the ROI of the initiative, and 2) list all the myriad benefits of improving customer data quality and hope to get enough buy-in to get the initiative funded. Neither of these work on their own.
In my experience, you need to lead with the second approach -- tell the "story" of customer data quality and get heads nodding. Then, follow up (in the same discussion) with the lowest hanging fruit -- what you can quantify that also has a significant payback. THEN, combine the two: "We can quantify *x*, which shows enough of a return to warrant undertaking the project, but there's a lot of other benefits that are extremely difficult/fuzzy to measure. So, 'x' is the baseline, and the initiative stands on its own based on that. BUT, the initiative bubbles to the top of the list when you accept that there are a slew of additional benefits that aren't worth trying to pin down specific measures on.
[...] “new kid” on the Information Quality blogging block, William Sharp. In his post “Begin at the End - Ensuring Data Quality Success” elegantly sums up one of the challenges in developing, presenting, and implementing [...]
ReplyDelete