Some weeks ago I wrote an "Open Book" on TheStreet.com in which I pointed out the contributions of poor data quality to the credit crisis and advised investors, regulators and company leaders to demand far better data.
As a reminder, poor quality data is one of the top two or three underlying causes of the credit crisis. Excessive greed is right up there as well and, unfortunately greed and poor data quality have uncanny ways of exacerbating one another. Lack of transparency and incorrect (sometimes falsified) data make it easier for the excessively greedy to proceed unchecked. And the greedy are strongly motivated to be opaque and to "manage the facts."
stimulated many discussions. While most people agree with my analysis, they also think I was too harsh. They point out that, in many cases, the data were "good enough." Income data on mortgages were "good enough" to make the loans, descriptions of complex products were "good enough" to sell them, and, until recently,
were "good enough" for banks to loan to one another.
These discussions have convinced me that I was not direct enough.
No one would argue, for example, that a brake disc that failed, killing a driver and her passengers, was "good enough" because it fit into wheel assembly. Nor would anyone argue that a surgery in which the patient died was "good enough" because it was a technical success.
Since the beginning of the Industrial Revolution (at least) quality has proven itself a winning long-term strategy. Anyone who doubts this point should compare the annual reports of
with their respective competitors. Companies like these, that embrace strategies of high-quality, enjoy lower costs and more satisfied customers. In contrast, of course, over the long-term, inattention to quality is a big loser. However, large and growing markets can cover up mediocre quality. Cell phone service is a good example.
Quality matters most in times of market turmoil. It is then that the full extent to which the curse of "good enough" reveals itself. The quality revolutions in manufacturing illustrate this point. Markets simply dried up for products that were "good enough" only a few months earlier. "Good enough" is not "one step below 'good'" or "two steps below 'excellent.'" It is the polar opposite.
Indeed, "good enough" is the enemy of continuous improvement, the only proven way to stay in sync with -- even
-- market changes.
We may well be at the onset of the quality revolution in
. If history is any guide, the revolution will be relatively short and brutal. Indeed, a "highly-leveraged quality revolution" may be even shorter and more brutal.
Some people asked me to provide more specific advice about what companies should do beyond "demand better" and "read my latest book,
Data Driven: Profiting from Your Most Important Business Asset
." Here, sharply condensed, are the five most important points:
1. Don't depend on regulation.
More regulation is coming. But more regulation is no more than a small part of the solution to the credit crisis, and I doubt it will save anyone in a quality revolution. Witness
. Developed in response to the
scandals, it was supposed to ensure that we could trust a company's balance sheet. From where I sit, Sarbox failed its first serious test, despite billions spent in compliance.
2. Focus on customer needs.
Managers and companies must follow the law, and they must meet accounting and other standards. But doing so is not enough. They must recognize that quality is in the eyes of customers. They must make sure they understand the needs of downstream users of financial data and/or reports. They must recognize that different customers have different needs (in this regard, a regulator is but one customer) and they must segment customers accordingly.
Conversely, they must act like customers and clarify their needs to those who supply them financial data.
3. Measure quality, publish the stats and improve.
Once companies understand customer needs, they must measure quality, as non-judgmentally as they can, against those needs. Some call this "gap analysis." It can be a sobering experience, as companies usually find they're not as good as they think. The real goal is to get over the emotional reactions and start making improvements. I find that companies with the courage to publish their results improve faster.
4. Create clear accountability at the source:
Perhaps the most important lesson in all of data quality management is "get it right the first time." It is simply too difficult, time-consuming and expensive to make corrections along the way. The only way to do get it right the first time is to hold sources accountable.
5. Manage end-to-end:
The hierarchical organizational structure has many virtues, clear lines of reporting and technical excellence among them. But for data it is anathema. Data born in one place lead fascinating lives, moving from one department to the next, lighting down to help one person complete an operation, a committee make a decision and a company report results.
Data spend most of their time "in the white space" of the organization chart, completely unmanaged. Organizations simply must identify and get their arms around their most important flows of data -- way back in the supplier base, across the various departments, and deeply into the customer base.
Some may dismiss these recommendations because they don't think we are really in the throes of a quality revolution. After all, there is no proven way to distinguish the early throes of a revolution from the latest round of market turmoil. If these people are correct, then those who don't implement these recommendations will recover and those who do will only reap incremental benefits. But what if they're wrong?
Thomas C. Redman ("The Data Doc") is president of Navesink Consulting Group in Little Silver, New Jersey. His latest book, "Data Driven: Profiting from Your Most Important Business Asset," was published by Harvard Business Press in September. Redman can be reached at