Monday, May 29, 2017

Another Step in Tableau's Evolution?

Tableau broke new ground when it was introduced. The best tool for simple, straightforward visual data analysis then, it remains unsurpassed in this space today, even as it's striving to be a player in the enterprise marketplace where, to paraphrase Christian Chabot, it can maximize its market potential.

Tableau always wanted to be more than a personal product. The early Polaris project at Stanford pursued the ability to visualize "large multi-dimensional databases [that] have become common in a variety of applications such as data warehousing and scientific computing." When introduced Tableau was a simple tool, but beautiful in its simple elegance, particularly compared to the options then available.

Even though it was predominantly useful only as a personal tool, when I bought my license in 2006 to use in my BI consulting work I had to purchase it from my assigned sales representative, who was keenly interested in leveraging my use of Tableau to penetrate my clients' organizations. And the price was exorbitant for a personal product in a world where Borland had introduced the model of highly functional low and moderately priced products that could be impulse-purchased.

Over the years Tableau added more features and functionality, often layering them on top of its base,  leveraging existing functionality rather than implementing new and better ways of doing things. Dashboards were a big step. Tableau Server arguably a bigger one. Always moving forward, Tableau addressed the concerns of "enterprise" people pointing out its shortcomings from their perspective.

The IPO was a big deal. It meant that Tableau had grown up in meaningful ways. Made it easier to raise capital and expand its visibility in the marketplace.

Making Gartner for the first time was huge. Now Tableau was visible to the top executives who look to Gartner to see what other top executives think, and Tableau was at a stroke a presence in the mindspace of enterprise business decision makers, a legitimate option for big corporate dollars to be spent on. Things were moving forward, life was good, especially for Tableau's founders and executives. They had made the world a better place, and reaped handsome rewards in return.

Then the haircut. The stock price dropped 50% in February, 2016. Moving into the enterprise marketplace was more difficult than disrupting a moribund personal tool space had been. Elegance and effectiveness of design had profoundly improved people's data-analytical lives in a burst, and Tableau became the darling of individual analysts. But big, serious, enterprise customers moved in response to a different set of forces. Conservative to the core, they needed to be convinced that Tableau could "be like" their existing heavily invested data processing tools and platforms. They didn't care that Tableau offered opportunities that could, if taken advantage of, dramatically change their entire approach to business data analysis, inverting it from a top-down industrial model to an organic model where the right data was available for use and analysis throughout the organization. Enterprise executives didn't want to do things better, they wanted to do the same things cheaper, and hopefully faster. Tableau announced that it was concentrating on high-value deals, with their longer lead times and associated complexity. The company that had succeeded because it sold individual licenses to people, making a difference in their lives, was now concentrating on big, impersonal deals. A different world indeed.

Tableau was in a vice. Its strengths had earned it a seat at the table, with a window of opportunity to demonstrate that it was an enterprise player like the big companies it had disrupted at the personal end of the spectrum. What to do?

Six months after the haircut Christian Chabot stepped aside as Tableau's leader after shepherding the company from frisky upstart to market leader, and then through the penetration of and acceptance in the enterprise market.

Tableau has been busy releasing new versions adding additional features that improve its enterprise offering. More data sources connected, more Server functionality, more stuff.

Tableau has not been busy improving its core ease of use, or of expanding the horizon of the things made simple and easy. It's almost as if Tableau believes that the 'low' end, the one where it stands out and is truly special, isn't the place to invest its time, energy, effort, and money.

Which brings is to the present, and the trigger of this post. According to reports Chris Stolte has sold of a bunch of Tableau stock, e.g.
  • Monday, May 15th, 75,000 shares at an average price of $61.69, for a total of $4,626,750
  • Monday, May 8th, 150,000 shares at an average price of $60.15, for a total of $9,022,500 
As of this information, Chris directly owns 78,805 shares valued at appx. $4,861,480

What does this mean? I have no idea. I don't know Chris, although I've met him a couple of times at Tableau functions.

But I'm worried. Worried that one of Tableau founders has stepped out of the direct line of effect, and another appears to be shedding himself of his investment in the company. Worried that Tableau is moving more purposefully towards concentrating on the enterprise market and maximizing its market potential, aka capitalization/shares value, and that this means that will move even farther from its roots as the best-ever tool for visual data analysis. Worried that the chances to revisit its core functionality, fill in the holes, correct the flaws, adjust the oddities, sand off the warts, improve its deficiencies, and extend its ease of use into new areas, are vanishing in the rear view mirror.

On the other hand, maybe Chris is looking for a new challenge. Maybe he's interesting in picking up again the mantle of champion of the visual analyst and work towards creating the next great visual data analytical tool, the one that is what Tableau could have become. If so, I've been working on some ideas.


  1. Chris,

    Sorry to go off-topic, but I thought commenting here would be the best chance to reach you. I've spent quite a few hours trying to get your TWB gem working. It looks like an incredible resource for auditing Tableau workbooks. Unfortunately, I'm not well versed in Ruby. Can you tell me- is it possible to get the TWB gem to audit Tableau 10 workbooks? If so, do I need a specific version of Ruby, and is there a preferred method to install the gem?

    Thanks for all your work on TWB,

    1. Hi Phil, sorry about the delay in replying.

      I've spent quite a lot of time recently upgrading TWB and the Ruby code for using it, am planning a series of posts on its use in practice, starting from very simple to pretty complicated.

      In the meantime, here are the basic steps to get up and running:
      - the current version of twb 1.1.0
      - have Ruby installed; I'm running ruby 2.4.1p111 (2017-03-22 revision 58053) [x64-mingw32] on Windows
      - make sure Ruby is part of your path (in Windows) so that it can be found when needed
      - install the TWB gem; with Ruby installed, at the command line use:
      > gem install twb
      - twb depends upon some other gems, some of these need also need to be installed separately:
      > gem install nokogiri
      > gem install rubyzip

      Once everything's installed, it's easily tested.
      From a command prompt, issue these commands (comments in parens)

      > irb --noecho
      (irb is the Ruby interpreter, )
      (--noecho suppresses its verbosity)

      irb(main)...> require 'twb'
      (makes the twb gem available)

      irb(main)...> twb = 'workbookname.twb'
      (assigns the new Workbook class to the twb variable)

      irb(main)...> puts
      (tell Ruby to print the workbook's name)
      irb(main)...> workbookname.twb

      irb(main)...> puts twb.version
      (tell Ruby to print the workbook's version)
      irb(main)...> 10.2

      I hope this gets you started.