Follow Datanami:
September 11, 2013

Tableau Throws a Brick at Traditional BI

Alex Woodie

Upstart business intelligence software vendors like Tableau Software have made a good living denigrating big BI vendors for some time. Tableau took the practice to new heights at its annual user conference this week, when its founders compared traditional BI to close-minded thinking, stone-age writing implements, and Mayan death chants.

Before we get to the zingers, let’s review the meat of Tableau CEO and co-founder Christian Chabot’s impassioned, 45-minute speech, which took place in front of 3,200 Tableau customers at the Gaylord National Resort and Convention Center in Washington D.C. and was televised live on the Web.

The gist of Chabot’s speech is that the greatest breakthroughs in the history of human thought–events like Albert Einstein’s creation of the general theory of relativity, Alexander Fleming’s discovery of antibiotics, Isaac Newton’s discovery of the laws of gravity, and David Stuart cracking of the Mayan code–can be attributed to a spark of creativity, or “thinking outside of the box” in today’s corporate-speak.

 ”If you studied the people who made the 100 most important discoveries of all time, you’ll find they all had something in common,” Chabot said during the keynote. “It’s not that they used data. They all used data too. It’s not that they know how to program computers better. Most of them didn’t know how to program. It’s not even that they’re smarter than their colleagues.

“Great minds do something special,” he continued. “Great minds combine–in an artful way–logic with intuition. Great insights are made by people who combine deduction with feel, structure with improvisation, the science with the arts. The great minds universally are jazz artists. They’re trained in the structure and language of their field, for sure. But they improvise in creative and unpredictable fashion, with an intuition that baffles their colleagues and has made all of them heroes.”

In other words, important breakthroughs do not occur by following a logical, repeatable, or formulaic process. That way may get you pretty far, but to make the leap to something truly novel and new, a discoverer must bring something intangible to the table, like resorting to his inner feelings or chasing down hunches.

Einstein was a true genius, and he worked in mysterious ways. “Einstein was famous for feeling his way to discoveries using his internal sense of things, like simplicity and beauty and harmony,” Chabot said. “He would famously throw away work after weeks of effort because it didn’t feel right. The man who discovered the most elemental laws of the universe said, ‘There’s no logical way to discover the elemental laws. There’s only the way of intuition.'”

In evening when the other scientists were putting their experiments away, Fleming would chide them for their orderliness, Chabot said. By comparison, his experiments were laid out in a haphazard manner so he could better see them. One morning, when Fleming came into his lab and noticed that a white halo surrounded a fungus he was growing in one of his experiments, he realized right away that this was a novel substance, which turned out to be penicillin.

Tableau CEO Christian Chabot speaks during a keynote at the Tableau user conference on September 9, 2013.

The Mayan code stumped many archaeologists who tried to understand it. It was so artistic and beautiful that some scientists gave up and figured it must just be there for decoration, Chabot said. What helped Stuart start to unravel its meaning was to try things that were never tried before. “There was a playfulness to the language that was as important as its structure, which is funny because David brought his own creative process,” he said.

When we do our best thinking, we do more than analyze and break things down. We do more than calculate and deduce, Chabot said. “That’s not the magic. We do that but that’s no the magic,” he said. “We chase hunches. We shift perspectives. We relate things spontaneously. We use intuition to feel our way through a problem. We improvise with information. In fact our best thinking is usually done when a set of observations lead to a question, which leads to an answer, which leads to a new question, new answer, new question, a hunch, and a new question and a new answer, and poof! You’ve got a spark. That’s how great thinking happens.”

In comparison, traditional BI tools work in the opposite manner, he said. “The dominant enterprise BI technology of our time…the technology used by 99 percent of all the companies and governments in the world, is designed to work in exact opposition to the way great thinking occurs. They’re architected and have their heritage in the age of computer that gives them a design that works in exact opposition to the way that great insights are known to develop in the human mind.”

Chabot then gave his impression of how traditional BI works. First, the software is installed. Then, the customer must go through “an elaborate requirements gathering phase where the questions are anticipated in advanced of the project, program an elaborate semantic layer that locks it all down, then give it to a specialist to try and make answers one at a time. Then the BI strategy becomes a series of change request sent from one department to another.”

This model does not provide the freedom of creativity that is necessary to make big discoveries, Chabot said. “I can’t ask an unanticipated in this model. I can’t shift perspectives on the fly. I can’t throw in new data. I can’t improvise. I can’t feel my way to an answer.  I can’t play jazz with this technology. I’d be luck to play a Mayan death chant for crying out loud!”

Later in the keynote, Tableau co-founder and chief development officer Chris Stolte took the stage with an interesting prop: a brick with a pencil taped to it. You could see the comparison coming from a mile away, but it turns out there is an interesting story behind it.

It turns out that the great computer pioneer Douglas Engelbart–who is credited with helping to create the mouse and hypertext, and who passed away this July at age 88–actually did an experiment in which he gave people bricks with pencils taped to them, and asked them to perform work. The experiment was recorded in Engelbart’s famous 1962 paper “Augmenting Human Intellect.”

Not surprisingly, the subjects in Engelbart’s experiment didn’t get much work done. Just lifting the brick up to write stuff left them exhausted an unable to think clearly. It led Engelbart to conclude that thinking is affected by the tools we’re given.

Tableau CDO Chris Stolte says using enterprise BI tools is like writing with a pencil attached to a brick.

“Enterprise BI is kind of a like a pencil with a brick attached to it,” Stolte said. “It’s dampening our creativity and it’s discouraging our understanding and our creativity.”

Stolte’s band of motley developers then took the stage to demonstrate some of the new features that are coming down the pike in the 8.1 release of Tableau’s product, due out later this year, and version 8.2, due out next year. Version 9 is a big release that’s scheduled for late 2014.

The biggest new feature in 8.1 will be integration of the R statistical language. “Starting with 8.1 you’ll be able to harness the power of R directly in 8.1. Model in R, visualize in Tableau, and enjoy a fluid experience of data analysis,” said Dav Lion, Tableau development manager.

 Other interesting tidbits include the inclusion of “box and whisker” plots in Tableau visualizations, which provide a clean indication of how far from the norm certain data points fall (the box) and how far out the outliers lie (the whiskers). The new release will also bring a cleaner interface for loading and cleaning data. “We want to make it so easy to get data into table that you’ll never hesitate to chase an idea or pursue a hypothesis,” Stolte said.