« How Playing Games on Your Phone Can Make You a Better Human | Main | You're On the Wrong Side of Your Face »

You Don't Need Big Data. You Need A Porsche. 

Posted on Tuesday, October 6, 2015 at 06:33PM by Registered CommenterLon Langston | Comments Off

As a perk of being on the Atlanta Aerotropolis Alliance board of directors, I was invited to the Grand Opening of the Porsche North American Headquarters building, on May 7th. The building is cool. The test track is cooler. It’s like a miniature road-racing course. It has a couple of straightaways and lots of turns and elevation changes.

I got to ride in two different Porsches, with two different professional drivers. Both drivers were incredibly good (obviously). The woman was better than the man, by the way. She navigated the course faster and exited the turns on better lines, requiring less correction before accelerating down the straightaways. It was like being on a roller coaster, with one big difference, the person in the car with me, behind the wheel, controlled my fate.

Big data is a popular catch phrase in the business, for now. There is an advantage to big data. It creates accurate statistical probability. But reams of data are useless. As Nate Sliver pointed out, in his May appearance on the Freakonomics Podcast, big data sets do not insure clarity, the opposite, in fact. The bigger the data set the more opportunities there are for statistical outliers. In a set with a billion data points, there are a lot of one-in-a-million occurrences. One-in-a-million occurrences are not what we want to base decisions on.

Besides, for most of us, in small and medium-sized businesses, big data is just an amorphous concept. What we care about is our data – small data. By “small,” I mean finite. There is an advantage with finite data, mostly that it’s finite. It is possible, in a small or medium-sized company, to identify every data point.

If you have done flow and process work, you’re probably thinking of examples which prove this wrong. Okay, I’ll concede that there is a level of magnification beyond which you begin to chase a dime with a dollar. Identifying data points, like every human activity, has an opportunity cost. My point is that a functionally-complete set of data points can be practically accumulated. The investment of time and money is worth the return.

I am married to a former math and science teacher, which is evidence that opposites attract. I did not learn much in school (half my fault, certainly, and half the one-size-fits-all education system’s fault), but no subject, for me, was worse than math. In college, I had to retake 2 classes, both math classes. I only passed Quantitative Methods because a math-geek friend took pity on me and tutored me.

So it was odd when spreadsheets entered the business world that I was drawn to them. I found that I like analyzing numbers. I like looking for trends in numbers. I like empirical metrics. At first, I didn’t like the fastidiousness necessary to collect accurate data, but I love the story accurate data tells so much that I became fastidious.

Having been in management in small, medium and large organizations, I know that decisions are based on intuition more than on data. One reason is lack of trust in the accuracy of the data. Another is that decisions have to be made fast. There isn’t time to collect and analyze a bunch of data. Opportunity will pass us by.

Great drivers use intuition to get the best out of the vehicle. A Porsche with a sucky driver isn’t going to perform and a Chevy Cavalier with a great driver isn’t going to perform. A great vehicle gives the driver the tools she needs to turn in a great driving performance. Data is a vehicle.
Good data is like a Porsche. A good driver layers her intuition on top of the vehicles capabilities. A good leader layers his intuition on top of accurate data.

Data cannot replace intuition. Businesses will fail, if decisions are based on data analytics, alone. Intuition is important. Good data just helps with good decisions.

So question 1 is how do we get Porsche-quality data and not Cavalier-quality data?

We get good data by creating a culture of attention to detail and by building accurate data collection into the normal routine, by making it standard operating procedure.

As we flew around a tight turn, with my face involuntarily pressed against the window by centrifugal force, I heard my driver explaining the features of the vehicle in an unhurried, conversational voice.

Question 2 is how do we get actionable data, fast enough to matter?

Urgency is a tricky thing. I was in the fresh produce industry. Our products were literally decreasing in quality every day that they sat in inventory. Sell it or smell it. Urgency was paramount.

If we don’t make the right decision, in a timely manner, opportunity will pass us by. Conversely though, if we do make the wrong decision, in a timely manner, misfortune will not pass us by. Making good decisions fast is all about preparation. If the Porsche driver had to search for the data she needed about the car’s performance, she would have to slow down or crash. Instead, the data is displayed on the dashboard and the heads up display. She knows right where to look, to get right what she needs, right when she needs it.

In order to make good decisions fast, prepare before time imposes constraints. Build a report that gives decision-makers access to the important stuff, at a glance, all on one screen. And get in the routine of looking at it daily.

A good data dashboard, with which all key decision-makers are familiar, is like the speedometer in a Porsche. You know right where to look, at the critical moment, to get the information you need.