Wednesday, February 20th, 2013
big research

Are you tired of “big data” yet?

The term has taken the business world by storm and is unavoidable as a topic of conversation among managers, technologists, and investors .  We all feel like we should have big data and do something with it.

As part of an ongoing series on equity research,the research puzzle | This PDF index is updated as new postings come out. here are some musings about big data in that context.  Whether you think that big data is a revolutionHarvard Business Review | This October 2012 article talks about the promise of big data. or “a marketing campaign, pure and simple,”Perceptual Edge | Stephen Few titled his piece “Big Data, Big Ruse. there are implications for investment processes from the explosion of data.

Most definitions of big data include “the three Vs” of volume, velocity, and variety.  Truth be told, that construct is now twelve years old, having originated at the META Group, since bought by Gartner.Gartner | This is the original PDF.  So, we’ve been dealing with these concepts for a while, but now the buzzwords are in widespread use.

There are a range of issues here for investment organizations.  Compliance and risk management can be improved by better organizing and analyzing the disparate data sources already at hand — although some more prosaic steps, like attacking spreadsheet errors,research puzzle pieces | This is from the third site dealing with “the research puzzle.”  It features short “pieces” about the business. might provide more bang for the buck.

The major big-data front in the equity realm has been high-frequency trading, with less feverish quantitative strategies also resting on predictive analytics.  The approaches of two decades ago were similar to today’s, it’s just that the dials for the three Vs have been turned up.  Let’s look elsewhere though.

If there is promise in using big data to gain economic advantage, equity investors should be investigating which firms can actually capture that advantage.  Cutting through the stories will be hard work (see Hewlett-Packard and Autonomy) and parsing the hype versus the reality will be reminiscent of the work on other such management fads and fancies over the years.

For those that actively manage portfolios in a qualitative manner, the big-data question remains the same as it has been for some time.  To what extent is your decision making data-driven?  (Or, to what extent should it be?)  Many firms allow great flexibility in how analysts and portfolio managers ultimately decide what to do after chewing through the data that is tossed around them.  But bits and bytes of information don’t make a process.

To take advantage of the opportunities, firms should not only rethink existing roles,the research puzzle | See, for example, “those darn analysts.” but the entirety of the investment function.  Structurally, data scientists and decision scientists may be mixed in with stock jockeys, making for a cultural clash.  The “how” of decision processes should morph if we truly are entering a different information environment.

As always, the leaders of investment organizations will determine whether the promise of all of this will be captured and the alpha revealed.  Which firms will be bold enough to lead after sniffing out these changes in the wind:

There will be more research firms that produce higher-frequency observations than the monthly and quarterly economic and corporate reports that we rely upon now.  There will be new partnerships that involve complementary sharing of information.  There will be investment organizations that purchase or team up with data aggregators and analyzers that seem far removed from the workings of the markets.  All of this will also bring new questions about privacy boundaries and regulatory constraints — and insider trading allegations too.  Perhaps some retailer will understand the gold mine of information it has accumulated and in essence create an internal investment fund to capitalize on it.  (I have no idea of the legal implications of that, but we should think expansively about the availability and use of data, since consumers seem willing to sign their information away at the drop of a hat.)

Consider a Bloomberg terminal today.  It can report a vast array of data, and it relays searches and graphs and favorites and messages from all over the investment ecosystem.  I haven’t read my Bloomberg agreement lately, so I can’t recite its policies.  For my purpose today, I don’t really care.  Imagine all of it was yours — the facts, the connections, and the behavioral clues.  What could you do with it?

I won’t argue that simple and sound techniques from years gone by are destined for the scrap heap as a result of changes in the information environment.  But “research” is bound to get bigger, providing opportunity and risk for those pioneers who try to capture the frontier, as well as those who stay behind.