Predicting Return on Investment from a data quality program (and many other business initiatives) is like predicting the weather. There is a myriad of factors, events and not fully understood processes that makes weather forecasting and making a business case for data quality a chaotic discipline. More on this in the posts Whether Weather Forecasting or Not and Miracle Food for Thought.
Also, when looking for data quality flaws, remember that all that glisters is not gold.
The classic business case for a data quality investment is avoiding sending duplicate letters and printed materials to the same individual. But this is only the first of at least 55 reasons to improve data quality that comes to my mind for improving data quality related to the single most frequent data quality issue around, which is duplicates (and unresolved hierarchies) in party master data – names and addresses.
One of these many other reasons is related to fraud. Here I have stumbled upon an example with big time ROI in identity resolution – 5 billion taxpayer Euros may have been saved if some identity resolution including automated fuzzy matching and real world checks was implemented.
But I think earnings and savings from better data quality may be much higher than that 
I agree that creating ROI calculations and business cases for data quality can be a bit tricky, especially if the people at the company (in my case the customer) are not used to creating business cases in general. Without business cases you’re stuck with gut feeling decisions, and I believe that pure data quality initiatives loose 9 times out of 10 when based upon gut feeling alone.
Here are my thoughts on calculating business cases for data quality.
1. First of all are you creating a business case for a complete and strategic data quality programme or just for an isolated business process? The first option requires an organization with a high level of data quality maturity.
2. By keeping them simple many small business cases can be chained together for an overall business case, for a strategic programme. My experience is that its not extremely hard to calculate these smaller business cases for tactical projects.
3. Don’t view a data quality business case differently than any other process improvement investment. At the end of the day you will save or earn money if you do something a bit better, no matter if its reducing duplicates or hiring more trained staff.
4. I guess it boils down to changing the management’s mindset. If they really care about and understand the true value of their data they will accept a more fuzzy business case for a strategic data quality programme.
After all great amounts of money are invested into security solutions to protect information. I doubt that all these investments are based on hard numbers and business cases.
I think its great that we DQ pros discuss this as we are pressed to constantly prove the value of data quality… again that’s my feeling!
Thanks Dario.
You prompted me to write a post about that simple business case for reducing duplicates: Returns from Investing in a Data Quality Tool .
The analogy between the business case estimates and weather predictions is right on. The difference is that quantitative/numeric meteorology is a science developed in the last 5 decades whereas the business case science for data quality has not been developed yet. I am trying… So far most data quality justifications are anecdote driven
From my perspective the next wave is in applications of Information Theory and possibly Information Economics to business case evaluations of data quality. It is not to replace the data quality anecdotes that can be powerful but rather to augment them
Thanks Larry. Information Economics and data quality is surely an exciting combination to watch in the future.