Friday, May 21, 2010

A Response to Stephen Few's "Douchebag" Detractor

We have made no secret about being fans of Stephen Few, linking to his blog and other work at least once a week.

Well a recent post by Few posted, in full, an argument made against his book by a claimed data visualization expert, Dennis Ross.

As fans of Few, of intellectual discussions about data visualization, and the importance of open opinions on these topics, we thought it was worth replying directly to Ross's points.

Ross's arguments against Few's book:

1. It's too academic.
Well the purpose is to put academic discussion behind visualization, and go beyond the "it's pretty!" analysis, right?

2. You can't tell clients they can't have what they want.
I don't think Few says that specifically, like, "I refuse to make a gauge". t's like survey, you give clients the best info you can, and then let them make their own decision. I totally disagree with just lying down and never giving a counter example. This was where the guy got insulting, and lost all credibility for me).

3. Bullet graphs suck, and are confusing to me.
In the right circumstance, they are good options. And just throwing out any chart form as too confusing is short-sighted, as there is almost always a perfect situation for any chart. Even a conical chart, always awful, could be useful to show how much of a cone has been constructed, right? And, if you only saw bullet graphs daily, they would make far more sense.
Perhaps I disagree with Few on the relative advantages and disadvantages of bullet graphs, pie charts, and bubble charts, though we all (I think) agree that it comes down to choosing the right graph for the data, and what you're trying to see.

4. Few's a douchebag because he hates Oracle/Xcelsius/etc.
Alongside this excellent point, some sub-arguments:

A. Tableau is ugly
So are you! See, I can make arguments that don't have reasoning too. And regardless of your opinion of the look of Tableau, it's more about the concept of visualizing complex multi-dimensional data. And visualizing it in a way that allows users to manipulate and adjust dynamically.

B. Oracle/Xcelsius are fine because they "can be improved, yes, but they are light years ahead of what we have been working with in the past 10 years."
Really? It's not OK to criticize because at least it's better than before? That's great. "Hey Iraqis, stop complaining, at least your government is better than it was 10 years ago!"

C. Oracle/Xcelsius "exist to serve customer or client needs, NOT Few’s design sensibilities."
Once again, this is the terrible POV that represents some of the worst areas of complacency in business. It is not an example of pig-headedness, or douchebaggery: it's actually a matter of personal responsibility. If I go into a restaurant and pick a sandwich off the menu that the server (a) thinks isn't very well made, (b) knows gets a lot of complaints, and (c) has a far superior, but similar, alternative in mind, she should tell me. In fact, the restaurant should probably stop including that sandwich on their menu in the first place.

But yes, if after all of that information is given, I still want my terrible sandwich, then the restaurant should probably make it for me.

(Speaking of, I have had several of my favorite meals taken off the menu of one of my favorite restaurants. I'm not sure what that says about me....)

**************************************************

As Few did in his posting, here are detractor Dennis Ross's sample dashboards.
Posted without comment.

We in this field of interest can freely disagree, but it is the arguments and defense that keep it strong. And it is only the strong arguments and defense which continue to advance the technologies we rely on (Tableau, Oracle, Xcelsius). Think of the amazing improvements between the oft-maligned default charting and color schemes of 2003 Excel, to the 2007 versions. And as people continue to discuss and debate Excel 2007's graphs, these will come in to making their future versions even better.

In short, by defining the entire visualization software discussions and debate as unnecessary, and those that ask for better as "douchebags", you are doing a great disservice.

**************************************************

This was a fantastic end quote from Few which I will undoubtedly use again:

"Keep giving your customers what they want, even when it doesn’t work, rather than taking responsibility as a consultant to add value. That may satisfy them for the moment, but it won’t help them in the least."

No comments:

Post a Comment