I sat in the audience at Waters USA 2015 watching my colleague Anthony Scriffignano, Dun & Bradstreet’s Chief Data Scientist, capture the undivided attention of the delegates with the “inconvenient truths” about decisions we make in today’s data-driven environment. It was intellectually powerful stuff – even for a very tuned-in audience of technology leaders and C-level executives at financial institutions who had gathered to learn more about emerging IT innovations and solutions they can adopt in their organizations for better decision making and optimization.
As Anthony spoke, I thought of how all of us – in both our personal and business lives – consume data at an alarming rate. Yet despite the consumption pace, we crave even more, causing us to run endlessly faster in order to simply keep up.
As a parent, my concern grows when witnessing my children become overly focused on the data that is available to them, linking them tightly to their devices, gobbling up whatever social or structured data is in their line of sight. Try as I might, I struggle to convince my high schooler that not every text message or Snapchat alert requires immediate attention. She believes that, whatever the problem, she can get the answer by seeking data from her phone.
It occurs to me that this lament is similar to that of most of today’s Wall Street Chief Data Officers. Their firms have mountains of data, more computing power than they could possibly need, and billions of dollars of compensation incentives driving them to perfect their correlations and analysis. The natural reaction to solving problems is often to get more data, when in fact sometimes it muddies the waters and obscures the causes.
I think of what my colleague Anthony stressed in his presentation, that many data managers run faster and faster to get just one millisecond ahead of the competition. Financial institutions continue to accumulate greater amounts of data, much of it unstructured data and social media related, to help gain intelligence on our markets. They process is ultra-fast. They hire the smartest quants in the world.
So with all this data and all the knowhow and computing power to run all these scenarios, what do they have now?
Our markets are generally more volatile and move unpredictably, while “flash crashes,” fraud and flawed market structure drive unprecedented regulation. Our institutions have volumes more data to analyze than ever before. But has that made our markets more stable? More trustworthy? Easier to understand? Are we more confident in our markets’ fairness or performance? Few believe so.
Participants’ reactions to the markets’ problems are disproportionately addressed by creating new regulation that restricts activity and often amounts to throwing the baby out with the bath water by curbing positive activity. Regulation may touch on some of the causes, however, it is often so broad that it masks the real problem, rather than pinpoint the source. There’s just too much data at hand, and without analyzing the relationships more closely, it has become impossible to target the specific data needed to identify and solve precise source issues…
So, despite all the data and intelligence, the current approach is missing something. Much of the industry is still performing data analysis like they always have (albeit faster and more effectively) – by crunching more data.
The real answer is to resist the urge to get more data for data’s sake.
The best CDOs are shifting their strategy to one that enables them to better recognize the importance of the relationship between the data, our flawed market structure, and ineffective risk management.
While the improvements in data analysis have created new opportunities for capital markets institutions, flaws remain in the system, illustrating that this data still has gaps that are not properly acknowledged. In part, this is the result of the industry’s effort to apply the most sophisticated algorithms and analytics to the most data it can find.
However, the inconvenient truth cited by Anthony is – simply – that more data is not necessarily better data. The answer lies in thinking differently about how the market consumes data and studies relationships in that data. This represents a marked change over the approach most CDOs took just a decade ago.
Now if I can only get my teenagers to think the same way…