Virgin on Business: Were Only Human

Data capture and data analysis can only be as reliable as the people doing the capturing and the analyzing.

June 27, 2017

Bill Virgin


Data is king. (Or is it data are king? Darn those Latin-derived plural nouns.) Data will cure disease, make us safer, win elections, run businesses and government more efficiently and effectively, and simultaneously allow us to sell more stuff to more people who are making smarter, more informed decisions about what theyre buying.

Data has always been with us, but we are living in the Golden Age of Data. With sensors embedded in every place and device, with networks to whisk that data to where it can be stored and studied, with sophisticated tools to mash together, parse, analyze and use data to make decisions, we are collecting more data, doing more with it and depending more on what it tells us than ever before.

The Seattle area tech cluster is a major beneficiary of Data Fever, with everyone from established players like Microsoft and Amazon to relatively new entrants like Tableau targeting facets of data capture, storage and analysis for their huge growth potential.

The data will set you free, Boeing and Ford alum Alan Mulally often told business executives, his own as well as others seeking wisdom. To hear the promises, thats the minimum to be expected from the power of data.

It truly sounds like a wonderful future, and dont we all wish we could live in it. Unfortunately for us as well as for the future and the data industry theres a gigantic problem with this glorious tomorrow that will render data no more effective at curing our ills than atomic energy, space travel, biotech or the personal computer were.

The problem: Humans are involved.

Not to be dismissive of an entire species, but whenever you find human involvement, theres your fatal flaw, the fly in the ointment, the bug in the system. Humans are irrational, emotional, illogical, inattentive, inconsistent, easily distracted, feckless, prone to whims. Thats on their good days.

And when humans apply those gifts and traits to the art and science of data collection and interpretation, mayhem will ensue. They wont collect the right data; if they do, theyll ignore or misinterpret what it is saying. Even if artificial intelligence winds up doing all the analysis and learning, itll be humans writing the algorithms into which the data is fed.

Think not? Consider recent human misadventures with data. The banking business is already awash in data about such subjects as mortgage finance credit quality, repayment history, default rates. The evidence of a looming catastrophe was there, in bank earnings reports, bond documents and foreclosure numbers, if anyone cared to look for it and read it correctly. A few sharp operators did. The rest of the industry didnt. A recession and huge losses in jobs and wealth resulted.

The Hillary Clinton campaign was as data driven as any campaign in American political history, one that could summon information about the electorate almost to the household level. Such was the reliance on the numbers that a top campaign executive would dismiss doubters including, according to one tell-all tome, Bill Clinton, who had won a few elections in his time with the data run counter to your anecdotes. But the campaign failed at executing one very simple arithmetic exercise: counting to 270.

In the early days of computing, the saying garbage in, garbage out arose as a warning that, no matter how fast or powerful the machine, it was only as good as the information it was given. That warning looms even larger today, as people put greater reliance on data to do the thinking for them. There will still be a place for the gut instinct, the hunch intuition and the lucky guess, especially if the answer to bad data faulty processes leading to mistaken conclusions is to stuff even more data into the system.

Monthly columnist BILL VIRGIN is the founder and owner of Northwest Newsletter Group, which publishes Washington Manufacturing Alert and Pacific Northwest Rail News.