Datadoodle

You are here: Home / Archives for definitions

definitions

“Tame” the data makers

by

I’ve heard of “taming data.” But the week before last at Strata I heard it in a new context: taming behavior.

Taming data has been “nichy,” as fellow TDWI writer Steve Swoyer puts it. He says, “It doubtless explains the etymology of, for example, Tamr.”

But Swoyer pushes on from there, as Swoyer knows how to do.

[Notice the] consonance between to wrangle and to tame. Both are grounded in the same metaphorical frame. Both are grounded in the same metaphor. This pre-conscious framing/understanding of the issue is more interesting than the stupid terms.

Former IBM sales engineer Lamont Lockwood, now the “Integration Expert,” sees two definitions. One is simple: to straighten and calm streaming data. “You don’t have time to fix it later,” he explains. “You need smart models to keep up.”

That leads to Lamont’s second, “nefarious” definition: taming the users who produce the data. “You’ll be trackable every day and every minute,” he muses, “like call-center workers….This is happening.”

Still a “tool” by any other name

by

A marketing manager I know stopped me in mid sentence. He didn’t want me to call his business intelligence product a “tool.”

Why? “It sounds small,” he said. But it is small, I pointed out. It’s smaller than many others in its space. It’s downloaded in under a minute and unpacks itself on a desktop in a few minutes.

But he waved that rationale away as if it were a fly, and I should have known. Marketing people, like the parents of gladiators, prefer their progeny to be perceived as big. Bigness casts dark shadows over competitors and conceals weakness. Industry insiders give big competitors good odds.

Users, though, have more immediate, personal concerns. They want something that feels good, works consistently, and adapts easily. This describes a “tool,” a label that should be taken as a compliment, not an insult.

To understand the value of good tools, read what farmer and essayist Wendell Berry writes about them. Over the years, he’s thought about them often, such as in his 1970s essay on the Marugg grass scythe.

It is the most satisfying hand tool that I have ever used. In tough grass it cuts a little less uniformly than the power scythe. In all other ways, in my opinion, it is a better tool because, it is light, it handles gracefully and comfortably even on steep ground, it is far less dangerous, it is quiet and makes no fumes, it is much more adaptable. In rank growth one narrows the cut and shortens the stroke. It always starts — provided the user will start. Aside from reasonable skill and care in use, there are no maintenance problems. It requires no fuel or oil. It runs on breakfast. Its cheaper to buy than most weed eaters and is cheaper to use than any other power mower. And best of all its good exercise.

I’d bet that everyone dreams, at least secretly, of software that matches the Marugg. Sadly, though, people with other agendas usually make the final decision — people whose careers depend on buying not tools but “solutions.” My friend the marketing manager has to appeal to those who write the checks. But I don’t care. I’ll keep saying “tool.”

Skinning “analytics,” the word

by

“Analytics,” the term, has been twisted so badly that Wayne Eckerson last month felt moved to rescue it with a definition. Rather, two definitions, possibly more.

One definition is capitalized, the other is not. What “analytics” might mean in italics, all caps, or underlined he doesn’t say.

Whatever the typography, Wayne just might have the stature to make it all stick. He’s been around the industry for nearly two decades, now as the TechTarget director of research and president of BI Leader Consulting. People know him, respect him, and like him.

The capital-A meaning takes the “macro perspective.” He says it’s “the processes, technologies, and best practices that turns data into information and knowledge that drives business decisions and actions.” The small-A version means “various technologies that business people use to analyze data.”

Referring to Tom Davenport’s use of “analytics” in his book titles instead of “business intelligence,” Wayne seems to imply that “analytics” should replace it elsewhere, too.

That suits me. “Analytics” does something “BI” can’t do. It throws light on the real point of the industry: making sense of data.

But Wayne’s proposal is doomed. No definition will stick that makes us refer to a dictionary before each use. I would still have to pause before dropping either one into a conversation, and that would probably be the same for most other people, I suspect. That kills it.

Am I the only slow learner around here? I asked for opinions from my modest network of data analysts. A reply came from just one of them (who asked for anonymity), far fewer than normal. That analyst emailed that he doesn’t care what “analytics” means. He added, “What is the deal with such pompous, elaborate definitions?”

Exactly. What is the deal?

The terms that stick do so in an instant. Tableau seems to have pulled it off with its word for visualizations, “viz.” It’s simple and sounds like it must have been picked up on “the street.” They also repeat it often in their blog, and a cadre of devoted users sing along.

Wayne muses toward the end of his post, “There’s more than one way to skin a cat.” Yes, sooner or later, we’ll come up with a best practice. But for now, this cat has run away, unskinned.