Month: June 2016

Bohemian Grove a la BI

The Bohemian Grove of the BI industry convenes for the fifteenth time in just three weeks. Naturally, you ask the obvious question: Are you serious? The Grove? A summit?

The answer begins with a fond recollection of the Grove. If you’ve never attended the Bohemian Grove yourself — I haven’t, though I’m sure I know someone who knows someone who’s waited on someone’s table who has — you may know of it as that century-old, mid-July pow-wow of leaders from big-iron industry, national politics, and old-time movie-making. Ronald Reagan slept there.

They ate, and they told stories, and they all went back to work a little more satisfied with the world as it was. Or not. Something in the stories coming through that fine Cuban cigar smoke might have stirred their hearts.

Read more

Getting to less: executive-level storytelling at Schwab

A key part of telling a story with data, says a Schwab executive I talked to last week, is thinking like a business person.

“You want to go for the high-level ‘aha!’” says John F. Carter, Schwab senior vice president of analytics and business insight, in San Francisco. Start the main conclusion, similar to news reports. “Less is more, but the right less. Executives don’t have time to get into the weeds.”

“The right less” has a long tail, where most of the data should go.

Less is also hard. The monthly summary comes after a long process of finding the story. Carter begins by meeting with a data analyst who has already sliced and diced a lot of data, such as new clients this month, new assets, new-client acquisitions. He’s turned it into preliminary conclusions and questions. A second meeting calls in units from marketing, business, and the field.

Read more

Data driven? No, you’re story driven, says German data scientist

Just when data seems to be approaching maturity as a touchstone in business, along comes a data scientist with doubts. Data is like the stone in stone soup, he says. It has not a single calorie of meaning.

“The beef is in data fiction,” says Joerg Blumtritt, CEO of Munich-based Datarella and consultant to a long string of large European and American clients in manufacturing and media. He comes to his observations after years of consulting and research.

“Fiction” in the same breath as data? Any kind of story at all frightens some data people, and now you talk about fiction! The standard fear seems to be that storytelling will distort data — as if data is bred and born pure like greenhouse cucumbers.

The real value of data, he says, is that it lets us talk about things we couldn’t talk about otherwise without resorting to subjective experience. It’s a touchstone.

It’s where you go from there that’s important, and that’s where things get messy. Meaning comes from context, which comes in stories and metaphors, he says. “I could tell you that the lawn outside my window is green. But that’s just data. It means nothing,” he says. The meaning of the green lawn comes only when he explains that that lawn will make the picnic at noon especially pleasant.

The question of the picnic’s quality, and the hopeful scenario in which people have a good time, determined the choice of data. With that scenario, other data was ignored. The day’s financial market data, the latest sports scores, and the organizer’s checking account balance were obviously irrelevant.

Data is filtered, selected, and rendered according to the needs of those who decide what’s relevant. One crisp example Joerg’s used is an MRI image of a knee. He explains in a 2014 presentation at the Content Strategy Forum in Frankfurt, Germany that the line representing the division of bone and tissue should be rendered differently for a surgeon than for a cardiologist. One needs a blurry line while the other needs a sharp line — two renderings of data reflecting the same knee.

There is so much data available now, he says, that you can prove almost any hypothesis. Big data makes it easy to prove just about any hypothesis. For example, Google data can produce many arbitrary subsets, all with highly significant differences. “What does that mean in the end?” he says. “Nothing.”

But where is all this going? What if, as he says, “the beef” is really in data fiction? It means, I think, that we might be free to play with data, to experiment freely and imaginatively. Insights, he seems to say, come from play and narrative instead of rigid “truth.”

Much of data analysis and predictive analysis is essentially fiction. Predictive analytics is about telling stories. “It’s not facts. It’s taking data and forming a picture from it that guides your decision.” It’s the same with personas, developed to guide marketing. Or credit scores to predict trustworthiness.

Fiction, he explained, is not random. Like data, it has to make sense by conforming to our experience. Data is the stage that constrains fiction, but fiction provides the images and drama.

“I hope that data will transcend from facts to fiction,” he says. “And I want to hear and tell the fairytale where we wake the sleeping beauty in data.”

I think what he’s saying boils down to this: Useful new understanding on which to base business decisions comes not from data. That understanding comes naturally and fluidly from the narratives that arise from data.

Data stories are what we do with data, giving it as little notice as we give our own breath. You hear those stories, though, if you listen carefully to people who analyze data. You hear metaphors, anecdotes, experiences, personas, scenarios and predictions.

The data kisses Joerg’s “sleeping beauty,” who awakens to inspire wise decisions.

Scenarios for a Qlik Sense data story

Real data storytellers know that data framed in a story is much more effective than data standing all alone by its naked self. They also know that storytelling is a giant that can lose much of its power when tied down to a vendor’s product.

We can applaud Qlik for respecting these facts. We can also applaud McKesson senior director of procurement analytics Jin Ro, who talked about storytelling last Thursday at the Qlik Virtual Forum. And let’s give an extra whistle for saving any mention of Qlik Sense for the end.

In his half hour keynote, he started with the case for stories: He recounted an experiment in which stories were shown to be far more memorable than data. Of subjects shown a series of presentations with data alone and with story, only five percent remembered any data but 95 percent remembered a story that had been told.

A story is also more persuasive, as shown by a different experiment: shown two pamphlets, the one showing data alone lost badly to a second one that told a story about a girl in need.

Eventually, he told a data story of his own. Back when he worked at Accenture, he helped a group move a group of suppliers from average 30-day net payment to 45 days and to calculate the benefits.

The group was using median averages, but Ro saw that that was the wrong way to look at it. Though median average do work in many cases, this wasn’t one of them. He persuaded the group to switch to spend-weighted averages. He did it using scenarios — that’s where the “story” comes in — in which he compared median and spend-weighted averages among several large suppliers.

Watch the video at the Qlik Virtual Forum on-demand page. Look for “Customer keynote: data, analytics, and storytelling.” (You’ll have to register.)

As a story, it seems to have been basic, though it’s hard to tell from his description. But it worked. And as a showcase for storytelling, his keynote was a hit.

What you can do: Try scenarios.